REASONING ABOUT DEVELOPMENT:

ESSAYS ON AMARTYA SEN’S CAPABILITY APPROACH

REDENEREN OVER ONTWIKKELING:

ESSAYS OVER AMARTYA SEN’S CAPABILITY APPROACH

THESIS

to obtain the degree of Doctor from the

Erasmus University Rotterdam
by command of the
rector magnificus

Prof.dr. H.G. Schmidt

and in accordance with the decision of the Doctorate Board.

The public defence shall be held on

Thursday 27 June 2013 at 13.30 hrs

by

THOMAS RODHAM WELLS

born in Zomba, Malawi.
 Doctoral Committee

Promoters:
Prof.dr. J.J. Vromen
Prof.dr. I. Robeyns

Other members:
Prof.dr. I. van Staveren
Dr. R.J.G. Claassen
Dr. G. van Oenen
 Contents

Introduction: How Should We Think about Poverty and Development? ..1

Chapter 1: An Outline of Sen’s Capability Approach .................................. 12
Chapter 2: Two Critiques of Sen ....................................................................... 39
Chapter 3: Judgement in Sen’s Capability Approach .................................. 59
Chapter 4. Sen’s Adaptive Preferences and Adam Smith’s Impartial
Spectator ................................................................................................................. 88
Chapter 5: Transformation Without Paternalism ....................................... 111
Chapter 6. Which Capabilities Matter for Social Justice? Democratic
Politics Versus Philosophy. .............................................................................. 145
Conclusion: Evaluation and Valuation .......................................................... 180
Bibliography ......................................................................................................... 187
Appendix A: Nederlandse Samenvatting ...................................................... 199
Appendix B: Curriculum Vitae......................................................................... 203
 Introduction: How Should We Think about Poverty and
Development?

Most of the world is enjoying the best standard of living, the greatest
wealth, and the greatest freedom to live valuable and meaningful lives of
any time in human history. But at the same time vast numbers of people
are living lives of stark deprivation which are made even more appalling
by the contrast. Indeed, it is the perspicuous contrast between the
quality of life open to some people but not others that both defines and
condemns poverty in the contemporary world: poverty is an
unnecessary state of deprivation that can and should be remedied. In
the poor world the general term for the removal of entrenched
deprivation is ‘development’.
Moreover, remediable deprivation exists not only in faraway places
with small economies, armed conflicts, or government repression, but
also within the rich world, with its homeless, jobless, sick, and socially
excluded or stigmatised. Deprivation can co-exist with great opulence.
For instance, even in a relatively wealthy country with an effective
welfare state, where urgent and straightforward human physiological
needs are largely met, there may be a great deal of absolutely real
‘relative poverty’, such as deprivation in the “social bases of self-
respect” (cf Rawls 1999). The rich world too seems to be in need of
development.
We are continually confronted with images of poverty and its
dramatic consequences for human lives on our television screens and
newspapers, and also with public debate about how to understand it and
what to do about it. But poverty is so pervasive that it seems to escape
human comprehension let alone solution. There are vast numbers of
people affected in many different contexts. Their poverty is apparent in
many different ways, from poor health to disabilities to lack of
opportunities or aspirations. The causes of poverty are likewise
numerous and include the interaction of physiological, environmental,
economic, social, and political factors.

The basic concern is with our capability to lead the kinds of lives we
have reason to value. (Sen 1999a, 285)

Over the last 30 years the Indian philosopher-economist Amartya

Sen has developed a distinctive normative approach to evaluating well-
being in terms of individuals’ freedom to achieve the kind of lives they
2

have reason to value, and development as the expansion of that

freedom. This freedom is analysed in terms of an individual’s
‘capability’ to achieve combinations of such intrinsically valuable ‘beings
and doings’ (‘functionings’) as being sufficiently nourished and freely
expressing one’s political views.
Development in this perspective is understood in liberationist terms:
of removing unfreedoms – ‘the domination of circumstances and chance
over individuals’- and of respecting and supporting individual agency
and societal self-determination to decide on and pursue the flourishing
life. Hence the title of Sen’s most influential book, Development as
Freedom (Sen 1999a). Those who argue for the moral priority of
increasing or redistributing wealth justify this because wealth is
generally useful for the freedom to live a flourishing life. Sen argues
that we should focus directly on achieving that goal rather than
‘fetishising’ one of the means to its achievement. Others argue that
happiness is the true measure and goal of objective well-being. Sen
argues that while happiness is obviously important it is not obvious that
it is the only aspect of life we have reason to value.
Sen’s capability approach has been enormously influential, and has
been taken up and developed by academics in many disciplines, as well
as NGO, governmental, and inter-governmental institutions concerned
with development and well-being (such as the United Nations
Development Programme). It may be helpful to analyse the diffusion of
Sen’s ideas in terms of the distinction between persuasion and
recruitment coined by Albert Hirschman (Hirschman 1992, 34). While
persuasion concerns the direct influence of new ideas and arguments on
those already working on a certain area, recruitment is concerned with
the indirect influence those ideas have by exciting the interest of
outsiders to come into the field. Perhaps because Sen has been seen as
excessively free with the standard theoretical structures for
understanding poverty and development it may be noted that Sen has
struggled to persuade many theorists, particularly in economics, to take
up his approach.1 But he has been very successful in recruiting across

1
Sen’s contributions across several different fields of economics were recognised by
the award of a Nobel Prize in 1998. Yet the direct influence of Sen’s capability
approach itself on mainstream economics has not been as great as one might expect.
In many development economics textbooks, for example, his earlier work on poverty
indexes and famines is given significant attention but capabilities are mentioned
superficially, dismissively or not at all. (E.g. “For Sen, poverty is not low well-being but
the inability to pursue well-being because of the lack of economic means” (Nafziger
2006, 178).) Its influence on orthodox welfare economics has been perhaps even
 3

inter-disciplinary boundaries, spawning a complex and sprawling

literature across academic disciplines as varied as ethics, sociology, and
even design and ICT. Martha Nussbaum, for example, then best known
as a classicist and Aristotelian philosopher, was drawn to Sen’s
capability approach by its non-utilitarian, non-Rawlsian features
(Nussbaum 1988) and collaborated with Sen at the World Institute of
Development Economics Research (WIDER) in Helsinki from 1987 to
1989.
Indeed, the attraction of Sen’s capability approach is not unrelated
to the fact that readers from many different backgrounds can see in it,
or project onto it, their own interests and concerns. This recruitment
effect can of course lead to fruitful inter-disciplinary research very
much in the spirit of Sen’s work. But it can also generate some
confusion (and frustration) as researchers from quite different
backgrounds bring quite different beliefs about what the capability
approach really is - a theory of justice, an account of agency, a non-
Welfarist welfare economics, a theory of sustainable development, and
so on - and therefore how it should be analysed and developed (cf
Robeyns 2005, 193–4). Freedom is thus at the heart of its appeal, but
also its difficulties in developing as a convincing coherent account.
Coherence and consistency are central virtues of analytical moral
philosophy, and so it is not surprising that the capability approach has
received a great deal of critical attention from philosophers. For
example, some have claimed that it is illiberal because it is not neutral
about the nature of the flourishing life (e.g. Sugden 2006); that it is
under-theorised (e.g. Pogge 2002); that it is excessively individualist (e.g.
Gore 1997); that its focus on freedom is ambiguous (e.g. Nussbaum
2003) or incoherent (e.g. Cohen 1993; Dowding 2006); that its
philosophical-economic understanding of agency is too abstract and
rationalistic (e.g. Giri 2000; Gasper 2002). Part of my thesis is concerned
with outlining my own understanding of the philosophical character of
Sen’s capability approach, often to the effect of showing that such
criticisms are misplaced if not mistaken. But one particular issue in the
philosophical approach to Sen’s writing on the capability approach bears
mentioning earlier because it relates to the approach I have taken in this
thesis.

slighter. While Sen’s work on social choice has been very influential, only a few welfare
economists have followed up his reformist agenda for a non-welfarist Welfare
Economics (for example Marc Fleurbaey, Erik Schokkaert, and Wiebke Kuklys).
4

Reading Sen as a Philosopher

Philosophers have shown interest in Sen’s practical philosophy right

from the beginning, but they have not always read him as charitably as
they might.2 The Oxford Companion to Philosophy defines the principle
of charity in interpretation as holding “that (other things being equal)
one's interpretation of another speaker's words should minimize the
ascription of false beliefs to that speaker” (Mackie 2005). It seems to me
that quite a number of Sen’s critics in academic philosophy have not
tried hard enough to set their own views temporarily aside and to
understand Sen’s work as far as possible in its own terms. They have
moved too quickly to identify his differences with their own views and
expectations as errors. Consequently some of the problems they identify
with Sen’s account are the products of misinterpretation rather than
bone fide problems for its credibility (cf Robeyns 2000). This can result
in the premature dismissal of Sen’s substantive claims (e.g. Roemer
1998; Pogge 2002; Dowding 2006; Sugden 2006) or excessively drastic
reconstructions (such as by Nussbaum, who rejects much of Sen’s
framework as too vague, worrying, or unworkable).3
I think that this unfortunate lapse may be partly explained (though
not justified) by Sen’s somewhat unorthodox style of doing philosophy,
which may in turn be related to his background in economics and its
rather more pragmatic orientation to theoretical work.4 Three aspects of
this unorthodox style seem to me particularly relevant to reading Sen as
a philosopher.

2
Exceptions to this, and somewhat responsible for the high profile Sen’s capability
approach has had in academic philosophy from the outset, include Bernard Williams, a
leading British moral philosopher, who provided comments on Sen’s second Tanner
Lecture on the capability approach (Williams 1989); Hilary Putnam who has written
extensively on Sen’s ‘Smithian economics’ (Putnam 2002; e.g. Putnam 2008); and John
Rawls, who taught a course on Social Justice at Harvard University with Amartya Sen
(and Kenneth Arrow) in 1968-9, and adapted his concept of primary goods somewhat
in the light of Sen’s “forceful” critique (Williams 1989, 168 fn. 8; Rawls 2005, 178–187
particularly 179 fn. 6). This influence can also be attributed to Sen’s early interest in
philosophical topics even at the start of his economics career (e.g. Sen 1966; Sen 1967)
and the close links he developed with leading Anglo-Saxon philosophers (including,
apart from those named above, Isaiah Berlin, Ronald Dworkin, Derek Parfit, Thomas
Scanlon, and Robert Nozick).
3
I discuss two such critiques, by Thomas Pogge and Nussbaum, in some detail in
chapter 2.
4
This small-p pragmatism is also apparent in Sen’s underlying concern with better-
worse relations rather than right-wrong dichotomies in his approach to rationality,
ethics, and political philosophy.
 5

First, Sen often bypasses traditional philosophical distinctions and

coins his own in order to focus on specific aspects relevant to what he
wants to say, which is often something different than the standard
vocabulary is most useful for. For example, freedom is a complex and
ambiguous idea but one influential distinction in the philosophical
literature (popularised particularly by Isaiah Berlin) is between the
concepts of positive and negative freedom. Yet Sen largely bypasses that
terminology and opts instead for a distinction between the “aspects” of
opportunity and process in freedom (Sen 2002a, 585–7) which is more
apt for the social choice perspective he wants to bring.5 Though it has
something in common with Berlin’s distinction it brings a different way
of looking at freedom.
The philosophical reader has to be aware of the unconventionality of
Sen’s terminology in order to avoid attributing positions to him that he
does not hold.6 I believe such a misreading accounts, among others, for
G.A. Cohen’s accusation that Sen distorts the proper meaning of
freedom as individual control in talking of such things as “freedom
from malaria” (Cohen 1993; Cohen 1994; cf Kaufman 2006); and Martha
Nussbaum’s rejection of Sen’s freedom talk because of her construal of
it as negative liberty and her association of that with neo-liberalism
(Nussbaum 2003, 44–7), a point I discuss further in chapter 2.
The second aspect of Sen’s unorthodox style is the limited and
pragmatic use he often makes of the terminology he coins. It often
seems intended to further a particular argumentative point by labelling
a relevant distinction, rather than to identify viable operative concepts
suitable for systematic theoretical analysis in the manner the analytical
philosopher is trained to expect. Failing to recognise this aspect of Sen’s
style can lead to confusion or frustration in the philosophical reader
and to the perception of Sen’s work as unphilosophical and in need of
reconstruction (or outright rejection).
Consider Sen on commitments, a ‘concept’ that has been taken up by
academic philosophers and has even had a special issue of Economics
5
Opportunity freedom is about access to valuable states (i.e. capability), while process
freedom is concerned at the systemic level with properties like fairness and at
personal level with properties like non-interference (Sen 2002b, 624).
6
And this is not only the case for philosophical terminology. For instance, Sen’s
carefully worded endorsement of “maximisation” in his account of rationality has
sometimes been misunderstood (e.g. Anderson 2001; Sen 2001, 57–58). Economists
often conflate “utility optimisation” with “maximising behaviour” but Sen distinguishes
the two. While optimisation requires choosing the best option, maximisation “only
requires choosing an alternative that is not judged to be worse than any other” (e.g.
Sen 1997a, 746).
6

and Philosophy devoted to it (21(1), 2005). This is the closest Sen came
to defining it in his famous Rational Fools paper:

Commitment is, of course, closely connected with one’s morals. But

moral this question is in a very broad sense, covering a variety of
influences from religious to political, from the ill-understood to the
well-argued (Sen 1977, 329).

As a positive definition of a philosophical concept this is distinctly

unsatisfactory. It is vague and ambiguous in exactly the way many
philosophical critics of Sen accuse him of being. Yet if one looks instead
at how Sen uses the term ‘commitment’ to draw a perspicuous contrast
to assumptions of egoism in economic theory it is as precise, clear, and
effective as one could wish a philosophical argument to be.

If the knowledge of torture of others makes you sick, it is a case of

sympathy; if it does not make you feel personally worse off, but you
think it is wrong and you are ready to do something to stop it, it is a
case of commitment. I do not wish to claim that the words chosen
have any very great merit, but the distinction is, I think, important.....
The characteristic of commitment with which I am most concerned
here is the fact that it drives a wedge between personal choice and
personal welfare, and much of traditional economic theory relies on
the identity of the two. (Sen 1977, 326, 329 emphases added)

It seems to me that philosophical readers are sometimes distracted

from considering what Sen is very clearly saying because they are overly
concerned with what exactly his terms mean. In this case a distinction
that pretty effectively demolishes a foundational tenet of orthodox
utility theory may be misread as proposing a unitary concept of
commitment. Without ever attempting to clarify the nature of
commitment as a concept, rather than a class of exceptions to
(extended) egoism, Sen goes on to show that his distinction does real
work, by opening a space for a broader understanding of rational choice
that better fits how humans actually behave as well as better
recognising human agency (e.g. Sen 2005, 10–12).
The third aspect of Sen’s style that needs to be stressed is the
philosophically unorthodox way that the ‘concepts’ he identifies relate
to each other. Many of these are negative rather than positive. That is,
like ‘commitment’, they are meant to show the insufficiency of a
particular theoretical perspective rather than the sufficiency of a new
operative concept. Yet even Sen’s positive conceptual apparatus seems
 7

intended not to show the right way of looking at the matter, but a better
one for certain purposes. Thus even central terms like ‘functionings’ and
‘capability’ may fairly be called philosophically obtuse: vaguely defined
and ambiguously related (for example, it’s not at all clear which literacy
is). They lack the robustness and suitability for operational analysis that
philosophers are used to. For example Sen begins his Dewey Lectures on
Well-being, Agency and Freedom thus,

The main aim of these lectures is to explore a moral approach that

sees persons from two different perspectives: well-being and agency.
Both the "well-being aspect" and the "agency aspect" of persons have
their own relevance in the assessment of states and actions. Each
aspect also yields a corresponding notion of freedom. (Sen 1985a,
169)

The relationship between well-being and agency has been a

particular source of puzzlement and strained reading by academic
philosophers. Nussbaum for example finds Sen’s distinction “puzzling”
and “obscure” and concludes that it is a “vestige of utilitarianism inside
Sen’s nonutilitarian project” that should be dropped (Nussbaum 2011a
Appendix B). Other philosophers have interpreted well-being and agency
as distinct operational concepts which are related in a ladder-like way:
first one achieves well-being, then one can have agency.
These ‘problems’ seem to follow from reading Sen as proposing a
kind of theory – unitary, pure, coherent – that he is not. The aspects of
well-being and agency he identifies are, as he calls them, aspects rather
than concepts proper. Thus unlike ‘proper’ philosophical concepts they
can overlap with each other in what Nussbaum finds a “puzzling” way.
They relate to ways of looking at a case, of which kinds of information it
is appropriate to consider in taking up a particular perspective, not a
claim about which concepts are correct. As Sen explains,

Although the agency aspect and the well-being aspect both are
important, they are important for quite different reasons. In one
perspective, a person is seen as a doer and a judge, whereas in the
other the same person is seen as a beneficiary whose interests and
advantages have to be considered. There is no way of reducing this
plural-information base into a monist one without losing something
of importance. (Sen 1985a, 208).

As an example, consider Mohandas Gandhi’s famous 1932 ‘fast unto

death’ against untouchability in general and the British government’s
8

plan for separate electoral constituencies for untouchables in particular.

This can be seen in terms of well-being freedom (choosing not to be
well-nourished); well-being achievement (being poorly nourished);
agency freedom (e.g. choosing to prioritise a moral concern – promoting
a more inclusive and egalitarian form of Hinduism - over his own well-
being); agency achievement (among others, a short term rapprochement
with his political rival, B. R. Ambedkar, who claimed to represent the
interests of India’s untouchables) (cf Lelyveld 2011, chap. 9). Each of the
aspects Sen identifies would seem to have some prima facie claim to
significance in an evaluation of Gandhi’s fast.
Sen often talks about the underlying complexity and ambiguity of
the ideas – like agency, freedom or well-being – that he is grappling with,
and he identifies “aspects” and “perspectives” which are helpfully
informative for different purposes. He really doesn’t seem to be trying
to identify what the correct account of e.g. freedom is, but rather to be
identifying better (or criticising unsatisfactory) ways of understanding
it. Indeed, the centrality of perspective has itself been an enduring
subject for explicit methodological reflection by Sen (e.g. Sen 1980; Sen
1983a; Sen 1983b; Sen 1993a), and is an inescapable feature of The Idea
of Justice (in the form of his interest in Adam Smith’s device of the
impartial spectator).
As I mentioned, Sen’s distinctive style of doing philosophy may have
much to do with his training and experience as an economist. Though I
won’t attempt to develop that analysis systematically, it seems to me
that Sen’s ‘perspectival’ approach is especially reminiscent of how the
empirical social scientist characteristically works. One tries out different
ways of viewing the complex social world in a way analogous to how an
astronomer may experiment with different polarising filters on a
telescope.7 One filter, for example, may make a certain class of stars
more prominent and thus make it easier to focus on and examine them.
(One may of course replace physical ‘filters’ with theoretical constructs
for observation, such as statistical models.)
This is an interactional approach to observation in that the observer
interacts with the world through experimenting with different points of
view by which to look at it, assessing these in terms of how helpfully
informative they are – what they add - for the observational purpose at
hand. Among other things this approach is comfortable with retaining

7
This analogy was suggested to me by Kevin Hoover, in a presentation about the
history and methodology of econometrics.
 9

multiple points of view as differently informative, and toggling between

them as circumstances require, while also being capable of identifying
and rejecting many (classes of) points of view as generally inadequately
informative. It is worth noting that this interactional feature of social
science research does not mean that the theories of observation
employed determine findings in themselves. The task of observation
here does not consist of the straightforward application of the single
best general theory of observation, but rather with exercising one’s
expert judgement about which theoretical resources to employ, and
how, in order to best understand this particular issue. (I take up this
issue of judgement in chapter 3.)
As I have tried to show with my examples (and will demonstrate
further in chapter 2), philosophers have not always recognised or
appreciated these distinctive features of Sen’s style, with sometimes
unfortunate results for their reading of him. That is not to say that
critical philosophical analysis of Sen’s implicit operative concepts, of
freedom for example, isn’t legitimate and important, and potentially
useful for understanding Sen’s claims and whether and how they work.
(In later chapters I myself pursue a more systematic understanding of
Sen’s device of the ‘impartial spectator’, the problem of ‘adaptive
preferences’, and the ‘exercise’ of social choice.) Nor that the theoretical
development of Sen’s distinctions into operational concepts cannot take
on a life of its own and produce a significant and insightful literature (as
the concept of ‘commitment’ has been developed within theoretical
accounts of rationality), even if it no longer bears a relationship with the
concerns of its supposed originator. But it seems to me that the
automatic application of such analysis has in practise sometimes missed
Sen’s actual concerns in favour of subsuming Sen’s terminology into the
author’s own theoretical tradition.
In this dissertation I have therefore tried to explore Sen’s work as far
as possible in its own terms. One of my aims was that my reading
should allow room for the breadth of Sen’s work, by paying him the
courtesy of assuming an essential continuity across his very extensive
publications - which include an unusual number of disciplines and
topics, from Indian economic development to rational choice and
personal identity - as relevant to a fuller understanding of his capability
approach.8 Of course this remains a philosophical thesis, focused on the

8
Sen himself frequently asserts such a continuity. For example, in his popular
synthesis of the capability approach, Development as Freedom (1999), he explicitly
10

examination and elaboration of Sen’s ideas (indeed, only chapter 5 is

directly concerned with human development as conventionally
understood). But it is founded on a charitable reading of Sen’s ideas and
arguments in light of his expressed concerns, rather on holding them to
account for their failings on a narrow analytical score-card. I do not
mean to dismiss such criticisms, or the theoretical traditions from
which they arise, since they may be both legitimate and in certain
contexts quite significant.9 But I want to focus on a more positive
analysis of Sen’s capability approach that tries to analyse how far it can
succeed in terms of what it intends to do and with the resources it has
available.

Outline of the Dissertation

Chapter 1 provides an overview of Sen’s capability approach in terms of

its theoretical development, structure, and applications, and also
outlines its theoretical development by various philosophers.
Chapter 2 addresses philosophical criticism of the
‘operationalisability’ of Sen’s work on the capability approach and
argues that much of it is misplaced. The arguments of two such critics -
Thomas Pogge and Martha Nussbaum – are examined in detail
Chapter 3 complements the negative analysis of chapter 2 – what
Sen’s capability approach is not – with a positive analysis of Sen’s
methodology of evaluation in terms of the account of judgement
developed by Samuel Fleischacker. When properly understood Sen’s
distinctive vices, as identified by orthodox philosophical critics, can be
understood as distinctive virtues. They are constitutive of how Sen’s
capability approach is supposed to work, as a contribution to the
evaluation of advantage rather than as a theory of justice.
Chapter 4 considers the challenge that ‘adaptive preferences’ pose
for the capability approach. The problem of adaption is of ‘valuational
neglect’, that people develop values reactively rather than
autonomously. It is a challenge because evaluating a person’s capability
requires determining two distinct aspects: their effective freedom to live
(i) the life they have reason to value (ii). I show that Adam Smith’s device

collates, in an accessible way, the central arguments and claims of many of his
publications in apparently disparate disciplines.
9
In debates with luck-egalitarians or theorists of fair compensation in welfare-
economics, for example, the lack of a normative account of individual responsibility in
the capability approach becomes significant (cf Robeyns 2005, 192).
 11

of the impartial spectator can provide a theoretical resource for

understanding, identifying, and remediating adaptive preferences.
Chapter 5 analyses the transformational implications of
development policy for personal identity and shows that the capability
approach, properly understood, respects individual agency in a way that
conventional approaches to development do not. Development can be
understood as transformational and in the interests of those concerned
only if people are treated as autonomous agents whose own valuation of
the life they have reason to value is central to the evaluation of
advantage and development. This understanding addresses standard
arguments against the paternalism supposedly embedded in the idea of
development, and links this response to the capability approach’s
democratic understanding of policy and society.
Chapter 6 analyses the role of democracy and philosophy in
determining which capabilities are important as a matter of social
justice (though, of course, justice is not the only perspective for which
the capability approach is relevant). I respond to the challenge of the
political philosophy gap in Sen’s work by outlining and defending his
pragmatic but optimistic ‘idea of democracy’.
12

Chapter 1: An Outline of Sen’s Capability Approach10

The capability approach is concerned with evaluating how well off

people are in terms of their capability to achieve the kind of lives they
have reason to value. A person’s capability to live a good life is defined
in terms of the value of the combinations of ‘beings and doings’ - like
being in good health and having loving relationships with others - to
which they have real access. This focus distinguishes it from more
established approaches to evaluating social welfare, such as
utilitarianism or resourcism, which focus exclusively on subjective well-
being or the availability of means for a good life, respectively.
The capability approach was first articulated by Amartya Sen in the
1980’s, and remains most closely associated with him. It has been
employed extensively in the context of human development, for
example by the United Nations Development Programme, as a broader
and deeper alternative to narrowly economic metrics like growth in GDP
per capita. Here ‘poverty’ is understood as deprivation in the capability
to live a good life, and ‘development’ is understood as capability
expansion.
Within academic philosophy the novel focus of the capability
approach has attracted a large number of scholars. It is seen to be
relevant for the normative evaluation of social arrangements beyond the
development context, for example for analysing gender inequality (e.g.
Robeyns 2003). It is also seen as providing foundations for normative
theorising, such as a capability theory of justice that would include an
explicit ‘metric’ (which capabilities are valuable) and ‘rule’ (how they are
to be distributed). Martha Nussbaum has provided the most influential
version of such a capability theory of justice, deriving from the
requirements of human dignity a list of central capabilities to be
incorporated into national constitutions and guaranteed to all up to a
certain threshold (Nussbaum 2000; Nussbaum 2011a).
This chapter outlines the development, structure, application, and
philosophical context of Sen’s capability approach. Section I explains its
evolution in relation to other ethical approaches (utilitarianism and
resourcism). Section II gives an account of its basic structure and
introduces the key aspects of evaluation and valuation. Section III
discusses some practical implications and applications. Section IV

10
This chapter is adapted from my Internet Encyclopedia of Philosophy article, Sen’s
Capability Approach (Wells 2012).
 13

provides an introduction to the academic philosophical literature on

Sen’s account by outlining various philosophical efforts to theorise the
capability approach, including but not limited to that of Nussbaum.

I. THEORETICAL DEVELOPMENT
Amartya Sen had an extensive background in development economics,
social choice theory (for which he received the 1998 Nobel Prize in
Economics), and philosophy before developing the capability approach
during the 1980s. This background is pertinent to understanding and
assessing Sen’s capability approach because of the complementarity
between Sen’s contributions to these different fields. Indeed Sen’s most
influential and comprehensive account of the capability approach -
Development as freedom (Sen 1999) – explicitly draws on and
synthesizes many of these particular, and often quite technical,
contributions.
Sen first introduced the concept of capability in his Tanner Lectures
on Equality of What? (Sen 1979) and went on to elaborate it in
subsequent publications during the 1980s and 1990s.11 Sen notes that
his approach has strong conceptual connections with Aristotle’s
understanding of human flourishing;12 and also with the work of Adam
Smith and Karl Marx, who, each in their own way, also discussed the
importance of functionings and capability for human well-being. For
example, Sen often cites Smith’s analysis of relative poverty in The
Wealth of Nations, in terms of how a country’s wealth and different
cultural norms affected which material goods were understood to be a
‘necessity’;13 and Marx’s foundational concern with “replacing the

11
Significant publications include a technical presentation in Commodities and
Capabilities (Sen 1985b); his second Tanner lectures; On the Standard of Living (Sen
1985c); his Dewey lectures, Well-Being, Agency and Freedom (Sen 1985a); Inequality
Reexamined (Sen 1992), and his ‘popular’ synthesis, Development as Freedom (Sen
1999a). Succinct overviews are Development as Capability Expansion (Sen 1989a) and
Capability and Well-being, in a volume co-edited with Martha Nussbaum (Sen 1993b).
The capability approach also plays a major role in The Idea of Justice (Sen 2009a)
12
Although Sen is careful to distinguish shared themes and concerns with Aristotle’s
approach from the differences, such as Aristotle’s commitment to a rather specific
view of the good life (Sen 1993b, 46–48). Aristotle’s accounts of flourishing and
political economy were the original basis of Nussbaum’s dignity-based alternative
capability theory (see particularly Nussbaum 1988).
13
“By necessaries I understand not only the commodities which are indispensably
necessary for the support of life, but whatever the custom of the country renders it
indecent for creditable people, even of the lowest order, to be without. A linen shirt,
for example, is, strictly speaking, not a necessary of life. The Greeks and Romans lived,
I suppose, very comfortably though they had no linen. But in the present times,
through the greater part of Europe, a creditable day-labourer would be ashamed to
14

domination of circumstances and chance over individuals by the

domination of individuals over chance and circumstances”.14
Sen characteristically begins accounts of the capability approach by
recapitulating the roles and limitations of alternative philosophical
accounts, particularly utilitarianism and resourcism. Sen argues that
whatever their particular strengths none of these provide an account of
well-being that is suitable as a general concept: they are all focused on
the wrong particular things (whether utility, liberty, commodities, or
‘primary goods’) and they are too narrowly focused (they exclude too
many important aspects from evaluation).

a. Sen’s Critique of Utilitarianism

Economics has a branch explicitly concerned with the evaluation of

social welfare (‘welfare economics’). Sen’s systematic criticism of the
form of utilitarianism behind welfare economics identifies and rejects
each of its three pillars: act consequentialism, welfarism, and sum-
ranking (Sen 1979a).15

i. Act-consequentialism

According to act consequentialism, actions should be assessed only in

terms of the goodness or badness of their consequences. This excludes
any consideration of the morality of the process by which consequences

appear in public without a linen shirt, the want of which would be supposed to denote
that disgraceful degree of poverty which, it is presumed, nobody can well fall into
without extreme bad conduct. Custom, in the same manner, has rendered leather
shoes a necessary of life in England. The poorest creditable person of either sex would
be ashamed to appear in public without them. In Scotland, custom has rendered them
a necessary of life to the lowest order of men; but not to the same order of women,
who may, without any discredit, walk about barefooted. In France they are necessaries
neither to men nor to women, the lowest rank of both sexes appearing there publicly,
without any discredit, sometimes in wooden shoes, and sometimes barefooted. Under
necessaries, therefore, I comprehend not only those things which nature, but those
things which the established rules of decency have rendered necessary to the lowest
rank of people.” (Smith 1776, V.2.148)
14
Karl Marx and Friedrich Engels, The German Ideology (1846), quoted in (Sen 1989a,
44).
15
See for further discussion of these issues (Sen 1979b; Sen and Williams 1982, pt.
introduction; Sen 1985a; Sen 1999a, 62). Although Sen sometimes distinguishes these
elements differently from their presentation here and gives them different labels
depending on the focus of his exposition, the essence of this critique remains the
same. Siddiq Osmani provides an excellent analysis of how “The Sen system of social
evaluation” can be understood as a critical response to these features of standard
utilitarianism (Osmani 2009).
 15

are brought about, for example whether it respects principles of fairness

or individual liberty. Sen argues instead for a ‘comprehensive
consequentialism’ which integrates the moral significance of both
consequences and principles (Sen 2000a). For example it matters not
only whether people have an equal capability to live a long life, but how
that equality is achieved. For largely biological reasons, women generally
live longer than men even under the same conditions. If the only thing
that mattered was achieving equality in the capability to live a long life
this fact suggests that health care provision should be biased in favour
of men. However, Sen argues, trying to achieve equality in this way
would override important moral claims of fairness which should be
included in a comprehensive evaluation (Sen 2002c; Sen 2006a).

ii. Welfarism

Welfarism (a term coined by Sen) is the view that goodness should be

assessed only in terms of subjective utility, whether that is interpreted
in terms of happiness or desire satisfaction.16 Sen argues that welfarism
exhibits both “valuational neglect” and “physical condition neglect” (Sen
1985b, 21). First, although welfarism is centrally concerned with how
people feel about their lives, it is only concerned with psychological
states (desires), not with people’s reflective valuations (whether they
would still endorse them if they thought about it and were aware of the
possibility of alternatives (Sen 1985b, chap. 3–4)). Second, because it is
concerned only with feelings it neglects information about physical
health, though this would seem obviously relevant to assessing well-
being.
Not only does subjective welfare not reliably track people’s actual
interests or even their urgent needs, it is also vulnerable to systematic
distortions that Sen calls ‘adaptive preferences’. As Sen puts it,

Our mental reactions to what we actually get and what we can

sensibly expect to get may frequently involve compromises with a
harsh reality. The destitute thrown into beggary, the vulnerable
landless labourer precariously surviving at the edge of subsistence,
the overworked domestic servant working round the clock, the

16
The common interpretation of utility in terms of revealed preferences is considered
a non-starter by Sen, not only because it makes a “heroic simplification” in assuming
that such a binary relation reflects a person’s well-being, but also because, pace John
Harsanyi, it doesn’t accommodate inter-personal comparisons of well-being (Sen
1985b, 18–9).
16

subdued and subjugated housewife reconciled to her role and her

fate, all tend to come to terms with their respective predicaments.
The deprivations are suppressed and muffled in the scale of utilities
(reflected by desire-fulfilment and happiness) by the necessity of
endurance in uneventful survival. (Sen 1985b, 21–22)

People can become so normalized to their conditions of material

deprivation and social injustice that they may claim to be entirely
satisfied with their lot. For example, following the Bengali famine of
1943 a health survey found that among widowers nearly half assessed
their health as “indifferent” (a synonym for “not at all good”) while no
widows did (Sen 1985b, 82–3). It is important to note the influence that
adaptive preferences can have on people’s physical condition. For
example, the perception of reality that follows from deeply entrenched
norms about what men and women deserve is an important aspect of
the reality of intra-family distribution of resources in many countries.17

iii. Sum Ranking

Sum-ranking focuses on maximizing the total amount of welfare in a

society without regard for how it is distributed, although this is
generally felt to be important by the individuals concerned. Sen argues,
following non-utilitarian philosophers such as Bernard Williams and
John Rawls, that evaluating social welfare by aggregating over
individuals in this way, “does not take seriously the distinction between
persons” (Rawls 1999, 24). But Sen also points out that individuals differ
in their ability to convert resources such as income into welfare. For
example a disabled person may need expensive medical and transport
equipment to achieve the same level of welfare. A society that tried to
maximize the total amount of welfare would distribute resources so that
the marginal increase in welfare from giving an extra dollar to any
person would be the same. Thus, in the all too likely event that people
do not all have the same ability to transform resources into welfare, the
straightforward sum-ranking approach will have profoundly
inegalitarian consequences in the distribution of both resources and
welfare (Sen 1979c, 215). Resources would be distributed away from the
sick and disabled to people who are more efficient at converting them
into utility.

17
I discuss the phenomenon of adaptive preferences and its challenge for the
capability approach in chapter 4.
 17

b Sen’s Critique of Resourcism

Resourcism is the view that the relevant metric of distributive justice or

inequality assessment is resources, and is characterised by a strong
commitment to neutrality about what constitutes the good life (i.e. with
what those resources may be used for). It therefore assesses how well
people are doing in terms of their possession of the general purpose
resources necessary for the construction of any particular good life.
Sen’s criticism of John Rawls’ influential account of the fair
distribution of primary goods stands in for a criticism of resourcist
approaches in general.18 Sen’s central argument is that resources should
not be the exclusive focus of concern for a fairness-based theory of
justice, even if, like Rawls’s primary goods, they are deliberately chosen
for their general usefulness to a good life. The reason is that this focus
excludes consideration of the variability in individuals’ actual abilities to
convert resources into valuable outcomes. In other words, two people
with the same vision of the good life and the same bundle of resources
may not be equally able to achieve that life, and so resourcists’
neutrality about the use to which resources are put is not as fair as they
believe it is (Sen 1990a, 118–9).
More specifically, Sen disputes Rawls’ argument that the principles
of justice should be worked out first for the ‘normal’ case, in terms of a
social contract conceived as a rational scheme for mutually
advantageous cooperation between people equally able to contribute to
society, and only later extended to ‘hard’ cases, such as of severe
disability. Sen notes that such cases are far from abnormal (accounting
for up to 10% of the global population (Sen 2009a, 258)) and argues that
excluding the severely disabled at the beginning risks building a
structure that excludes them permanently.19 The general problem is that
such accounts ‘fetishize’ resources as the embodiment of advantage,
rather than focusing on the relationship between resources and people. 20

18
Sen considers that his critique applies generally to resourcist approaches, including
for example that of Ronald Dworkin (see e.g. Sen 1990a, 115)
19
“Human diversity is no secondary complication (to be ignored, or to be introduced
‘later on'; it is a fundamental aspect of our interest in equality.” (Sen 1992, xi)
20
Sen often quotes Aristotle’s Nicomachean Ethics (book I section 5) on this point:
“Wealth is evidently not the good we are seeking; for it is merely useful and for the
sake of something else.” Nevertheless Sen acknowledges that although the distribution
of resources should not be the direct concern in evaluating how well people are doing,
it is very relevant to considerations of procedural fairness (Sen 1989a, 52).
18

II. CORE CONCEPTS AND STRUCTURE

a. Functionings and Capability

Sen argues that the evaluation of advantage should consider what

people are actually able to be and do. The commodities or wealth people
have (resources) or their mental reactions (utility) are an inappropriate
focus because they provide only limited or indirect information about
how well a life is going. Sen illustrates his point with the example of a
standard bicycle (Sen 1985b, 10). This has the characteristics of
‘transportation’ but whether it will actually provide transportation will
depend on the characteristics of those who try to use it: it might be
considered a generally useful tool for most people to extend their
mobility, but it obviously will not do that for a person without legs. Even
if that person, by some quirk, finds the bicycle delightful, we should
nevertheless be able to note within our evaluative system that she still
lacks transportation. Nor does this positive mental reaction show that
the same person would not appreciate transportation if it were really
available to her.
The capability approach focuses directly on the quality of life that
individuals are actually able to achieve. This quality of life is analyzed in
terms of the core concepts of ‘functionings’ and ‘capability’ (Sen 1993b,
31).
Functionings are states of ‘being and doing’ - such as being well-
nourished, having shelter, and so on – and should be distinguished from
the commodities employed to achieve them (as ‘bicycling’ is
distinguishable from ‘possessing a bike’).
Capability refers to the set of functionings combinations that a
person has effective access to. In evaluating advantage, the value of a
person’s capability is her effective freedom to choose between different
functioning combinations – between different kinds of life - that she has
reason to value. (In later work, Sen refers to ‘capabilities’ in the plural
(or even ‘freedoms’) instead of a single capability set, and this is also
common in the wider capability literature. This allows analysis to focus
on sets of functionings related to particular aspects of life, for example
the ‘capabilities’ of literacy, health, or political freedom.)
 19

Figure 1. Outline of the core relationships in the capability approach

Resources Capability set Functionings Utility

(Characteristics of (functionings achieved (subjective well-
goods available) combinations available) being)

Personal ‘utilization function’

(depending on interpersonal differences:
physiology, and physical and social
environment)

Figure 1 outlines the core relationships of the capability approach

and how they relate to the main alternative approaches focused on
resources and utility. Resources (such as a bicycle) are considered as an
input, but their value depends upon individuals’ ability to convert them
into valuable functionings (such as bicycling), which depends for
example on their personal physiology (such as health), social norms, and
physical environment (such as road quality). An individual’s capability
set is the set of functioning combinations that an individual has real
access to. Achieved functionings are those they actually select. For
example, an individual’s capability set may include access to different
functionings relating to mobility, such as walking, bicycling, taking a
public bus, and so on, while the functioning they actually select to get to
work on a particular day may be the public bus. Utility is considered
both as an output and as a functioning. Utility is an output because
what people choose to do and be naturally affects their subjective well-
being (for example, bicycling to work on a sunny day may be rather
more enjoyable than on a rainy day). However Sen also considers
subjective well-being – feeling happy – as a valuable functioning in its
own right and so incorporates it directly into the capability framework.21
Sen argues that the correct space for evaluating advantage is an
individual’s capability to live a life they have reason to value, and not,
for example, their resource wealth or subjective well-being. But merely
21
Note though that some capability scholars argue for an objective definition of
capabilities that excludes happiness. For example, in considering capabilities as a
metric of justice, Elizabeth Anderson argues that only an objective metric can deal
with adaptive preferences, be publicly transparent, and fit the basic requirement of
justice as posing second personal normative claims (Anderson 2010, 84–6). This is an
example of how employing the concept of capabilities in theorising about justice may
be seen to impose restrictions on its informational basis.
20

identifying this space does not say much about how we should go about
evaluating advantage. I think it is helpful here to draw out an implicit
distinction in Sen’s writing between the operations of valuation and
evaluation. Valuation concerns the assignment of value to certain
objects; evaluation concerns the assessment of whether such objects are
achieved.

b. Valuation: Which Functionings/Capabilities Matter?

In order to begin to evaluate how people are performing in terms of

capability, we first need to determine which functionings matter for the
good life and how much, or at least specify a procedure for determining
this. I call this logically prior operation valuation. One way of addressing
the problem is to specify a list of the constituents of the flourishing
human life on philosophical grounds (Nussbaum makes minimal access
to this the basis for her capability theory of justice). Sen rejects this
approach because he argues that it denies the relevance of the values
people may come to have and the role of democracy (Sen 2004a).
Philosophers and social scientists may provide helpful ideas and
arguments such as the concept of capability itself - but the legitimate
source of decisions about the nature of the life one has ‘reason to value’
must be the people concerned. In particular, when using capability
approach for public policy purposes (for example, as the basis for
justifying a policy intervention and the criteria for evaluating its
success), methodological decisions about which functionings to focus on
and their weighting must be open to public scrutiny by those concerned
(Sen 1999a, 78–79). This is because these decisions are the basis for a
social choice about how that society should be changed, and it is
illegitimate to make such social choices without a process that can
include the views of those concerned. Sen therefore proposes a social
choice exercise approach to valuation, involving both public reasoning
and democratic procedures of decision-making. (I return to this in
chapters 5 and 6.)
One reason that social scientists and philosophers are so keen to
specify a list is that it can be used to objectively score how well people
(or societies) are doing: by ranking all the different constituents of the
flourishing life with respect to each other it would permit
 21

straightforward measurement of individual advantage.22 Sen’s social

choice exercise is unlikely to produce collective agreement on a
complete ranking of different functionings, if only because of what
Rawls called “the burdens of judgment” (Rawls 2005 II.§2). But Sen
argues that substantial action-guiding agreement is possible (Sen 1999a,
253–4). Firstly different valuational perspectives may ‘intersect’ to reach
partial agreements about some issues, though by way of different
arguments. Secondly such agreements may be extended by introducing
‘ranges’ of weights rather than cardinal numbers. For example, if there
are four conflicting views about the relative weight to be attached to
literacy vis-à-vis health, of 1/2, 1/3, 1/4 and 1/5, that contains an implicit
agreement that the relative weight on education should not exceed 1/2,
nor fall below 1/5, so having one unit of literacy and two of health would
obviously be better than having two units of literacy and one of health
(Sen 1992, 46–49).
Sen does suggest that in many cases a sub-set of crucially important
capabilities associated with basic needs may be relatively easily
identified and agreed upon as urgent moral and political priorities.
These ‘basic capabilities’, such as education, health, nutrition, and
shelter up to minimally adequate levels, do not exhaust the resources of
the capability approach, only the easy agreement on what counts as
being “scandalously deprived” (Sen 1993b, 40–42). They may be
particularly helpful in assessing the extent and nature of poverty in
developing countries. However, taking a basic capability route has
implications for how the exercise of evaluating individuals’ capability
can proceed, since any evaluative exercise based on it can only
determine how well people’s lives are going in terms of that very limited
selection of dimensions.

22
Although complete ranking (an index) is not necessary. For example, Nussbaum’s list,
which is a proposal for the constitutional principles of a just society rather than for
evaluating advantage, identifies 10 centrally important capabilities but doesn’t weight
them with respect to each other. Nevertheless one can use such a list to tell how well
someone is doing (or how well a society is doing at meeting its social justice
obligations) in terms of their cumulative shortfall from the minimum threshold. “My
central project is to work out the grou[n]ding for basic political principles to which all
nations should be held by their citizens; but an ancillary and related project is to map
out the space within which comparisons of quality of life across nations can most
revealingly be made (Nussbaum 2000, 116).”
22

c. Evaluation: What Capability do People Have?

Evaluating capability is a second order exercise concerned with mapping

the set of valuable functionings people have real access to. Since it takes
the value of functionings as given, its conclusions will reflect the
normative choices made in the valuation stage, and will be limited by
those choices in its focus and precision.
Assessing capability is more informationally demanding than other
accounts of advantage since it not only takes a much broader view of
what well-being achievement consists in but also tries to assess the
freedom people actually have to access high quality options. This is not
a purely procedural matter of adding up the number of options
available, since evaluation requires making judgements about their
significance to actual people’s lives. For example, the option to purchase
a 10th brand of washing powder has a rather different significance than
the option to vote in democratic elections (cf Williams 1989, 98; Sen
1989b, 108–9). Nor is evaluation limited to choices. For example Sen
argues that the eradication of malaria from an area enhances the
capability of individuals living there even though it doesn’t increase the
number of choices those individuals have (since they don’t have the
‘option’ to live in a malarial area anymore) (Sen 1993b, 43–4). Because
the value of a capability set represents a person’s effective freedom to
live a valuable life in terms of the value of the functionings available to
that individual, when the available functionings are improved, so is the
person’s effective freedom.
The capability approach in principle allows a very wide range of
dimensions of advantage to be positively evaluated (‘what capabilities
does this person have?’). This allows an open diagnostic approach to
what is going well or badly in people’s lives that can be used to reveal
unexpected shortfalls or successes in different dimensions, without
aggregating them all together into one number. The informational focus
can be tightened depending on the purpose of the evaluation exercise
and relevant valuational and informational constraints. For example, if
the approach is limited to considering ‘basic capabilities’ then the
assessment is limited to a narrower range of dimensions and attempts
to assess deprivation - the shortfall from the minimal thresholds of
those capabilities - which will exclude evaluation of how well the lives of
those above the threshold are going.
 23

d. Determinants of Capability

As well as being concerned with how well people’s lives are going, the
capability approach can be used to examine the underlying
determinants of the relationship between people and commodities, and
thus play a role in explaining poverty and advantage. These
determinants include (Sen 1999a, 70–1):

Individual physiology, such as the variations associated with illnesses,

disability, age, and gender. In order to achieve the same functionings,
people may have particular needs for non-standard commodities (such
as prosthetics for a physical disability), or they may need more of the
standard commodities (such as additional food in the case of intestinal
parasites). (Note that some of these disadvantages, such as blindness,
may not be fully ‘correctable’ even with tailored assistance.)

Local environment diversities, such as climate, epidemiology, and

pollution. These can impose particular costs such as heating or clothing
requirements.

Variations in social conditions, such as the provision of public services

like education and security, and the nature of community relationships,
such as across class or ethnic divisions.

Differences in relational perspectives, such as the conventions and

customs that determine the commodity requirements of expected
standards of behaviour and consumption. As a result, relative income
poverty in a rich community may translate into absolute poverty in the
space of capability. For example, local requirements for ‘the ability to
appear in public without shame’ may vary widely.

Distribution within the family, such as the distributional rules within a

household that determine the allocation of food and health-care
between children and adults, males and females.

The diagnosis of capability failures, or significant interpersonal

variations in capability, directs attention to the causal pathways
responsible. Note that many of these interpersonal variations will also
influence individuals’ abilities to access resources to begin with. For
24

example disabled people often have more expensive requirements to

achieve the same capabilities, such as mobility, while at the same time
they also have greater difficulty earning income in the first place
(Kuklys 2005). The scope for uncovering such interactive effects is
important to the capability approach’s appeal. Judging people’s
advantage in terms of capability is justified not only because how well
people’s lives are going should be of direct moral concern (right object),
but also because doing so generates insightful perspectives (better
methodology). In the following section I briefly outline some of these.

III. IMPLICATIONS FOR PRACTICE

Sen’s capability approach has numerous implications for practise, in
areas as diverse as thinking about social justice or the empirical
measurement of poverty. This brief overview considers several of these:
treating individuals as agents; the incorporation of consequentialist
analysis; the global-local character of capabilities; disaggregated
evaluation; the capacity for perspicuous contrasts; and the improvement
of statistics.
Sen argues that expanding freedom, or capability, is both “the
primary end” and “the principal means” of development (Sen 1999a, xii).
Freedom is central to development because the effective freedom to live
a life one has reason to value is intrinsically valuable and thus the best
evaluative dimension of progress. But agency freedom is also the most
effective means for development since it not only directly enhances
well-being, but also enhances people’s ability to help themselves and to
improve their world (Sen 1999a, 18).
One implication of this is that the reasoned evaluations of the
individuals concerned should be central to judgements of social welfare
and policy action. Other accounts tend to treat deprived individuals as
“patients” with given characteristics who are to be the passive recipients
of rather technocratic development schemes. In contrast, Sen argues for
recovering a classical sense of the “agent”, as “someone who acts and
brings about change, and whose achievements can be judged in terms of
her own values and objectives” (Sen 1999a, 19).
Considering individuals as agents also focuses attention on the
importance of issues like incentives and real-world behavioural norms
for the social realisation of well-intentioned interventions. It can, for
example, reveal the need for planning to address the unintended but
quite foreseeable consequences of policy interventions. This, Sen
 25

believes, is an aspect often missed by moral theorists, especially those

working in what Sen calls the “transcendental institutionalist” approach
to theorising about justice (such as Rawls), who often make unrealistic
assumptions, such as of full compliance.23
Sen’s capability approach combines a concern with identifying
intrinsically valuable functionings with consequentialism about their
realisation. This relates directly to his long-standing claim that the
“engineering” approach to normative analysis (found particularly in
economics) “of emphasizing and pursuing logistic issues of
interdependence and interconnections” can complement the analysis of
moral philosophical arguments about the intrinsic importance of
various considerations (Sen 2007, 78). If one has an interest in actually
bringing about a world in which more of the things you consider
valuable are realised, then one needs to pay systematic attention to the
contingent empirical causal relations that are relevant to its realisation.
Such consequentialist analysis is not only concerned with identifying
trade-offs. Sen claims that some capabilities are instrumentally
interconnected in positive ways, and that some unfreedoms and
injustices are particularly interconnected in negative ways. This
consequentialist analysis can indicate particular priorities for public
action. For example, increased and widely shared educational
advancement (part of “social opportunity”) can increase economic
productivity; contribute to a more equal distribution of aggregate
national income; enhance the conversion of commodities into other
valuable functionings by individuals; and support people’s intelligent
decision-making about the kind of life they want to live (i.e. practical
reason) (Sen 1989a, 55).
This interconnectedness implies that it is not only directly desirable
but also often most effective for development to strive to enhance
multiple capabilities simultaneously. Thus, the five fundamental
capabilities Sen chose to focus on in Development as Freedom - the
rather opaquely named political freedoms, economic facilities, social
23
This is not only an issue for moral philosophers. George DeMartino notes that
economists advising policy makers routinely employ the “maxi-max” principle in
selecting and justifying particular policies or projects: out of the set of possible
policies choose the one with the best possible potential outcome (DeMartino 2011,
144–153). The fact that such policies may also have the greatest risks of failing and
causing significant harm to those who are supposed to benefit is neglected in the
selection process and not communicated to those whose lives are at stake. Since the
possibility of failure is hardly considered, there is also typically little attention paid to
ongoing monitoring of policies to spot problems early, or preparation of contingency
plans.
26

opportunities, transparency guarantees, and protective security - were

particularly chosen for their empirically supported positive instrumental
linkages to each other and to other valuable freedoms (Sen 1999a, 10).
One of the advantages of the concept of a capability like being well-
nourished or literacy is its definition can abstract from particular
circumstances even though its realization depends on specific local
requirements. Thus, the same fundamental capability to be well-
nourished can be compared for different people in different parts of the
world, even though for each of them it may require different amounts
and kinds of food depending on one’s age, state of health, dietary
restrictions, and so on. This gives the capability approach a ‘global-local’
character, applicable across political, economic, and cultural borders.
For example, Sen points out that being relatively income poor in a
wealthy society can entail absolute poverty in some important
capabilities, because they may require more resources to achieve (Sen
1983c). For example, the capability for employment may require more
years of education in a more affluent society.
The global-local character of capabilities can also be used to focus
on how their requirements are affected by social norms, thus placing
those social norms in the spotlight. Many capabilities have underlying
requirements that vary strongly with social circumstances (although
others, such as for adequate nourishment, may vary less). For instance,
the ability to appear in public without shame is a capability that people
around the world generally value (and have reason to value), but its
resource requirements vary significantly according to cultural norms
from society to society and for different groups within each society
(with respect to gender, class, and ethnicity). Presently in Saudi Arabia
for example, women must have the company of a close male relative to
appear in public, and require a chauffeur and private car to move
between private spaces (since they are not permitted to use public
transport or drive a car themselves).
Strictly speaking the capability approach leaves open whether such
‘expensive’ capabilities, if considered important enough to be
guaranteed by society as a matter of justice, should be met by making
more resources available to those who need them (subsidized cars and
chauffeurs), or by revising the relevant social norms. The capability
approach only identifies such capability failures and diagnoses their
causes. However, if there is general agreement (arrived at through an
exercise of social choice, as Sen suggests) that such capabilities should
 27

be equally guaranteed for all, there is a clear basis for identifying the
inequitable demands of such social norms as the direct cause of relative
deprivation and thus for criticising them as inconsistent with the spirit
of such a guarantee.
The capability approach takes a multi-dimensional ‘disaggregated’
view of advantage (Sen 1999a, 76–8). Often it may seem that people are
generally well-off, yet a closer analysis reveals that this ‘all-things-
considered’ judgement conceals surprising shortfalls in particular
capabilities, such as the sporting icon who can’t read. Capability analysis
rejects the presumption that unusual achievement in some dimensions
necessarily compensates for shortfalls in others. From a justice
perspective, the capability approach’s relevance here is to argue that if
people are falling short on a particular capability that has been
collectively agreed to be a significant one, then justice would require
addressing the shortfall itself if at all possible, rather than offering
compensation in some other form, such as increased income.
Capability evaluation is informationally demanding and its precision
is limited by the degree of agreement about which functionings are
valuable. However, Sen has shown that even where only elementary
evaluation of quite basic capabilities is possible (for example life-
expectancy or literacy levels), this can still provide much more, and
more relevant, action-guiding information than the standard
alternatives, such as GDP statistics.24 For example, Sen’s own empirical
research with Sunil Sengupta involved weighing the children in two
Indian villages. Their results showed that the richer village had almost
identical female under-nourishment levels as the poorer one (the
additional food seemed to have been almost entirely distributed to the
male children) (Sen and Sengupta 1983; Sen 1985b, chap. Appendix B).
This deprivation amidst relative plenty would have been missed in a
resource based survey, which generally cannot examine the distribution
of food consumption within a household.25

24
In addition, it is worth noting that the standard alternatives have their own
measurement problems. For example, although the concept of GDP is straightforward
enough, measuring it accurately is very demanding, especially in poorer countries.
Ghana’s GDP was recently revised upwards overnight by some 60% following a review
of the underlying methodology (Jerven 2013, 26–8).
25
This exercise only looked at the achievement of a single functioning (nourishment)
rather than a fuller evaluation of capability sets. In the practical application of the
capability approach to evaluation, there is a necessary trade-off between “relevance
and usability” (Sen 1985c, 27). Sen argues that, given the real-world constraints on
data, one should be pragmatic about how to employ the foundational idea of capability
in measurement. The important thing is that metrics determined by “usability”
28

In particular, by making perspicuous contrasts between successes

and failures the capability approach can direct political and public
attention to neglected dimensions of human well-being. For example,
countries with similar levels of wealth can have dramatically different
levels of aggregate achievement - and inequality - on such non-
controversially important dimensions as longevity and literacy. And,
vice versa, countries with very small economies can sometimes score as
highly on these dimensions as much richer ones (as figure 2 shows with
respect to South Africa and The Philippines). This demonstrates both
the limitations of relying exclusively on economic metrics for evaluating
development, and the fact that national wealth does not pose a rigid
constraint on such achievements (i.e. that GNP is not destiny). Such
contrasts are easily politicised in the form of the pointed question, Why
can’t we do as well as them?

Figure 2. Perspicuous contrasts: The Philippines does more with less

(Data from the 2010 UNDP Human Development Report26)

Philippines South Africa

Gross National $ 4,002 $9,812
Income per capita27
Life expectancy 72.3 52
(years)
Mean years of 8.7 8.2
schooling

Part of the capability approach’s achievement has been in promoting

the systematic collection and use of much more data on functionings
and capabilities so that more nuanced evaluation becomes possible. The
new Multidimensional Poverty Index (MPI) is a good example of this. 28

considerations should not be mistaken for what is ultimately of “relevance” (Sen

1999a, 81–5).
26
Since 1990 the United Nations Human Development Programme has issued annual
reports based on the Human Development Index [HDI] developed by Mahbub ul Haq
with Amartya Sen to compete directly with the crude but popular GDP per capita
figures. (The HDI is discussed further in chapter 2.)
27
At purchasing power parity, i.e. official exchange rates adjusted to allow for the
differences in the cost of goods and services between countries.
28
The MPI was developed by researchers at the Oxford Poverty and Human
Development Initiative (OPHI) and is included in the United Nations Development
Programme’s Human Development Reports from 2010 onwards. For an overview of its
methodology and applications see Alkire and Santos (2010a; 2010b).
 29

The MPI is based on newly available data sets from household surveys
and is designed to give a more nuanced picture of the nature and
intensity of poverty at the individual level by looking at combinations of
particular deprivations within the accepted standard three dimensions
of education, health, and the standard of living.
By looking directly at important functionings at the individual level,
the MPI provides a quite different picture of international and regional
poverty than income poverty (‘headcounts’ below an income poverty line
of, for example, $1.25 per day). More people are MPI poor (1.7 billion)
than income poor at the $1.25 level (1.3 billion); moreover some
countries with high income poverty have relatively low MPI poverty (for
example Tanzania), and vice versa (e.g. Ethiopia and Pakistan). This is
because the conversion of income into the functionings and key services
covered by the MPI depends on various contingent variables - such as
the quality, accessibility, and price of such goods as clean water - which
vary between countries especially in relation to their approach to the
public provisioning of goods. The MPI also provides a more nuanced
view of the geographical and social distribution, depth, intensity, and
character of poverty. Different countries and regions are revealed to
have different characteristic patterns of deprivations, which call for
particular policy responses. The MPI also allows the success of policy
interventions to be directly and fairly immediately assessed.29

IV. THEORETICAL DEVELOPMENT WITHIN ACADEMIC PHILOSOPHY

Sen’s capability approach has attracted a great deal of engagement from
academic philosophers, and a number of those sympathetic to Sen’s
foundational concern with the moral significance of the capability space
have introduced theoretical elaborations of his account.30 Although
Nussbaum’s capability theory of justice is the best known of these (and
will therefore be the subject of much critical engagement in this thesis),
in this section I also want to indicate the wider range of such work,

29
It should be noted, however, that the MPI is limited by the available data, collected
for other purposes, which preclude consideration of other relevant dimensions such as
work, safety and empowerment. Additionally, because the MPI is designed to focus on
severe deprivation, and therefore employs quite concrete indicators of deprivation,
such as, literally, whether a household has a non-concrete floor, it necessarily has a
‘low ceiling’ and is rather less applicable for discriminating the existence, depth,
intensity and nature of poverty in middle and high income countries.
30
Although the underspecification of Sen’s original account seems to be not only quite
deliberate on Sen’s part, but also integral to what he is trying to accomplish, and
therefore not necessarily a problem to be overcome (as I argue in chapter 3).
30

which, in contrast to Nussbaum’s, is generally concerned with

complementing rather than displacing Sen’s account.31 Some theoretical
accounts are primarily concerned with operationalising the evaluative
dimension of the capability approach: the assessment of quality of life,
well-being and human development (Robeyns; Alkire). Others focus on
developing a capability based ‘theory of justice’ in the spirit of its
concerns (Anderson; Alexander; Nussbaum).

a. Evaluation
i. Ingrid Robeyns: Generating Lists for Quality of Life Research in the
Social Sciences

In a methodology paper, Selecting Capabilities for Quality of Life

Measurement (Robeyns 2005), Ingrid Robeyns argues against a “strong
critique” of Sen’s capability approach: that it fails to specify which
capabilities are intrinsically valuable and this makes it useless for
normative evaluation (a critique most closely associated with Nussbaum
(e.g. Nussbaum 2003)). Attempting to develop a single all-purpose list of
capabilities would be incompatible with Sen’s concern to advance a
general approach to evaluation rather than a theory of either welfare or
rights (and his concern that, particularly where public action is
concerned, those concerned be involved as agents in determining the
relevant selection of valuable capabilities). Exactly because of its
“intrinsic underspecification”, Sen’s framework can be (and has been)
applied in many different forms, academic disciplines, levels, and scales,
and for different purposes (cf Robeyns 2006).
However, Robeyns does endorse a “weak critique”: “that we do need
some systematic methodological reasoning on how .... selection could be
done” (Robeyns 2005, 195). An important reason for this is that any
evaluation of advantage in terms of capability has to be based on
methodological choices: the selection of some capabilities rather than
others, how to specify these, and how to weight them in the overall
evaluation. There is a risk of “selection biases” in how this is done,
either reflecting mistakes (such as due to an insufficient grounding in
the appropriate academic literature) or preconceptions relating to
researchers’ social positioning or worldview (for example, Eurocentric or
31
Sen’s notes the “constitutive plurality” of the capability approach, that it “does not
lead to one particular theory of valuation (but defines instead a class of such theories
within a general motivational structure) (Sen 1985c, 27).”
 31

androcentric preconceptions). To address this, Robeyns proposes a

standardised procedural methodology to the selection of capabilities for
quality of life measurement. Robeyns’ five criteria are (Robeyns 2005,
205–6):

(1) Explicit formulation (All proposed list elements should be explicit,

so they can be discussed and debated);
(2) Methodological justification (The method of generating the list
should be made explicit so it can be scrutinized);
(3) Sensitivity to context (The level of abstraction of the list should be
appropriate to its purposes, whether for philosophical, legal, political, or
social discussion);
(4) Different levels of generality (If the list is intended for empirical
application or public policy then it should be drawn up in two distinct
stages, first an ideal stage and then a pragmatic one that reflects
perhaps temporary feasibility constraints on information and
resources);
(5) Exhaustion and non-reduction (The list should include all
important elements and those elements should not be reducible to
others (though they may overlap)).

Robeyns argues that her criteria provide a “check and balance” that
minimise the problem of selection biases and thus support the
epistemic, academic, and political legitimacy of empirical evaluations of
capability. She has used this methodology to select suitable dimensions
for conceptualising and assessing gender inequality in Western societies
in terms of capabilities, which she then applied in a survey of the
findings of existing empirical studies (Robeyns 2003).

ii. Sabina Alkire: A Participatory Approach to Evaluating Capability

Expansion

In her book, Valuing Freedoms (Alkire 2005), Sabina Alkire develops a

philosophically grounded framework for the participatory valuation and
evaluation of development projects in terms of capability enhancement
which incorporates the “principle of subsidiarity” that “the most local
agent(s) capable of making a choice should make it” (143). This
framework allows her to go beyond standard cost-benefit analyses of
development projects in financial terms to investigate which capabilities
32

that the people concerned have reason to value are enhanced and by
how much.
Alkire’s approach has 2 stages of evaluation. First, a theoretical one-
off stage in which ‘philosophers’ employ practical reason to reflexively
identify the basic domains or categories of value. Second, a local
participatory phase in which members of a social group deliberate, with
the aid of a facilitator, about what their needs are and what, and how,
they would like to do about them (with the basic categories employed as
prompts to ensure that all main dimensions of value are discussed).
For the first, philosophical, stage Alkire proposes an adaptation of
the practical reasoning approach of John Finnis to identify the basic
dimensions of human well-being by asking iteratively, ‘why do I/others
do what we do?’ until one comes to recognize the basic reasons for
which no further reasoned justification can be given. This method is
intended to yield substantive and objective descriptions of the
fundamental, non-hierarchically ordered, dimensions of human
flourishing, while allowing the content and relative importance of these
dimensions to be specified in a participatory process according to a
particular group’s historical, cultural, and personal values. The
intrinsically important dimensions identified by this method are: Life;
Knowledge; Play; Aesthetic experience; Sociability; Practical
reasonableness; Religion.
One of the advantages Alkire claims for her approach is its ability to
elicit what the people whose lives are the subject of development
projects really consider valuable, which may sometimes surprise
external planners and observers. Her use of the participatory approach
for assessing NGO fieldwork in Pakistan showed for example that even
the very poor can and do reasonably value other things than material
well-being, such as religion and social participation. Alkire’s approach
thus goes some way to realizing Sen’s foundational concern for
respecting agency in the practice of development.

b. Justice
i. Elizabeth Anderson: Justice as Equal Capability of Democratic
Citizenship

In an influential paper, What is the Point of Equality? (Anderson 1999)

Elizabeth Anderson argues against the dominant luck egalitarian
approach to egalitarianism in which the focus is on the redistribution of
 33

resources with regard to moral desert. She argues that egalitarian justice
is fundamentally political, about ending social oppression and
constructing a community in which people relate to each other as equals
(Anderson 1999, 288–9). She combines this fundamental concern with
“the better way to understand freedom” proposed by Sen (316), to
generate a partial theory of justice focused on equal capability of
democratic citizenship.

Negatively, people are entitled to whatever capabilities are necessary

to enable them to avoid or escape entanglement in oppressive social
relationships. Positively, they are entitled to whatever capabilities
are necessary for functioning as an equal citizen in a democratic
state. (Anderson 1999, 317)

Although Anderson’s primary concern is for equality in the

particular dimension of democratic citizenship, she notes that this has
extensive egalitarian implications for how society as a whole should be
organised. That is because other capabilities - such as relating to health,
education, personal autonomy and self-respect, and economic fairness -
are required as supporting conditions to realize truly equal citizenship.
For example, the economy would have to be re-engineered so that
unskilled work is no longer often rewarded with low-status and low
income, and infrastructure re-designed to fit the needs of the disabled
(rather than assuming the perspective of the able-bodied as ‘normal’)
(Anderson 1999, 334). However, Anderson points out that there is a cut-
off built into her egalitarian theory since it does not seek comprehensive
equality in the capability space but is focused on the political
dimension. The political relevance of additional functionings is finite
and satiable. Adequate education and food are required, for example,
but democratic equality doesn’t require PhDs or gourmet cuisine
(Anderson 1999, 329).

ii. John Alexander: Capability as Freedom from Domination

In his book, Capabilities and Social Justice (Alexander 2008), John

Alexander has proposed a capability theory based on a republican
understanding of the importance of freedom as non-domination. He
argues that the capability approach’s concern with people’s ‘real
freedom’ sets it outside and against the standard liberal egalitarian
theory of justice framework which understands freedom as the absence
of constraints. But he argues that the capability approach should go
34

further to elaborate this commitment to real freedom in Republican

terms. In this perspective it is not only important that one be able to
achieve certain functionings, such as mobility, but whether one’s
achievement of these are conditional on the favour or goodwill of other
people or are independently guaranteed by one’s own rights and
powers.
Capability is standardly understood as mapping one’s range of
choices over valuable functionings regardless of their content. For
example the ability of a physically disabled but socially well-connected
person to travel outside whenever she wants by arranging the help of
friends, family and voluntary organizations. In addition the republican
perspective requires that her capability for mobility should be
independent of context. This might take the form of a guaranteed legal
right to government assistance on demand, or by the provision of her
own specially adapted self-drive vehicle. Otherwise she may be said to
still be deprived since her capability is not robust with respect to the
whims of others.32
Alexander also suggests that domination should be integrated into
capability evaluation because domination is often a significant cause of
capability deprivation.33 It is no coincidence that the people who are
most capability deprived are often the poorest and weakest in society,
and as a result also vulnerable to yet further deprivation. This emphasis
on freedom from domination also gives a strong normative orientation
to the capability approach’s evaluation of the causes of capability
failure: some causes are simply unacceptable, such as social norms
restricting women’s freedom of movement and employment, and should
be removed rather than mitigated.

32
Philip Pettit (Pettit 2001) has also suggested a republican interpretation of Sen’s
capability approach. This example is taken from Sen’s response to Pettit in which Sen
accepted that the republican discrimination between types of freedom could be helpful
in capability evaluation, but that employing it did not require constitutionally
emending the capability approach in republican freedom terms (Sen 2001, 52–56). Sen
noted further that “We live in a world in which being completely independent of the
help and goodwill of others may be particularly difficult to achieve, and sometimes
may not even be the most important thing to achieve (56).”
33
Although domination is not the only cause of social oppression, which can often be
the result of institutions and norms rather than the conscious and intentional
behaviour of individual agents, as pointed out by Sharon Krause (Krause 2013).
 35

iii. Martha Nussbaum: A Multi-Dimensional Capability Theory of

Justice

Martha Nussbaum has contributed the most influential and

systematically developed capability theory of justice.34 Nussbaum aims
to provide a partial theory of justice based on dignity, a list of
fundamental capabilities, and a threshold. Her account is motivated by a
concept of dignity which she originally linked to flourishing in the
Aristotelian sense but now claims is the suitable object of a Rawlsian
overlapping consensus.35 She argues that human dignity requires the
securing to all of 10 central capabilities up to a minimal threshold36:

1. Life
2. Bodily Health
3. Bodily Integrity
4. Senses, Imagination, and Thought
5. Emotions
6. Practical Reason
7. Affiliation [social relationships]
8. [Relations to] Other Species
9. Play
10. Control Over One’s Environment

Nussbaum’s account is a partial rather than a complete theory of

just distribution because it is compatible with different accounts of
distribution about the threshold (Nussbaum 2000, 12). The threshold is
a sufficientarian principle of justice that specifies the minimum levels of
functioning in each dimension that a just society should guarantee to all
citizens access.
Universal access to the items on her list is required by human
dignity, Nussbaum argues, but this does not mean that a life lacking in

34
Key texts in the development of Nussbaum’s capability theory include her
development of an Aristotelian approach (Nussbaum 1988; Nussbaum 1993) and her
move over the 1990s to a Rawlsian style partial theory of justice based on an
overlapping consensus (Nussbaum 2000; Nussbaum 2003; Nussbaum 2011a).
35
“By ‘overlapping consensus’ I mean what John Rawls means: that people may sign on
to this conception as the freestanding moral core of a political conception, without
accepting any particular metaphysical view of the world, any particular comprehensive
ethical or religious view, or even any particular view of the person or of human
nature”. (Nussbaum 2000, 76)
36
For the full specifications, which have changed slightly over time, see Creating
Capabilities (Nussbaum 2011a, 33–4).
36

any of these, whether from external deprivation or individual choice, is a

less than human life. Nevertheless, the difference between choice and
deprivation is significant. If someone lacks access to any of these
capabilities, for example to be well-nourished (part of “bodily health”),
that reflects a failure by society to respect her human dignity. If
someone chooses not to take up her opportunity to be well-nourished,
for example to adopt an ascetic life-style or fast for religious reasons at
the expense of her bodily health, respecting that choice is also an aspect
of respecting her dignity.
Nussbaum suggests that her list, together with the precise location
of the threshold, should be democratically debated and incorporated
into national constitutional guarantees, international human rights
legislation and international development policy. In keeping with its
commitment to political liberalism, the components of Nussbaum’s list
have what may be called a ‘thick-vague’ character, in that while she
makes a universal claim for their substantive importance to any human
life, their definition is vague enough to allow their specification in
multiple ways that reflect the values, histories, and special
circumstances of particular political societies.37 Nevertheless, because
each capability is equally centrally important a shortfall in any area is
significant in itself, and thus the scope for governments to make trade-
offs between them, for example on the basis of quantitative cost-benefit
analysis, is limited.
Nussbaum’s account is particularly significant because its
similarities to Sen’s work has led to a general perception (outside
specialised readers) that she and Sen are working on the same project,
when in fact there are important differences. In addition, because
Nussbaum has developed her capability theory in a deliberately
philosophical way, demanding conceptual coherence and completeness,
many readers have concluded that she has taken the crude beginnings
of Sen’s capability approach and made it properly philosophical.38 This,

37
For example, Nussbaum suggests that freedom of speech (part of the capability for
affiliation) can be specified differently in law in the USA and Germany, because of their
different histories, without endangering the fundamental capability (Nussbaum 2004,
198). However, as White and Deneulin have noted, Nussbaum imposes quite strict and
ambitious limits about the permissible range of specifications, for example by entirely
rejecting gender-differentiation (White and Deneulin 2009, 255).
38
Aside from disagreeing that the straightforward application of the methodology of
analytical moral philosophy is the best way to do justice to Sen’s work and ideas, I am
also far from convinced that the substance of Nussbaum’s arguments is in keeping
with its form (cf Claassen and Düwell 2012). For example, the supposedly core concept
of dignity is invoked repeatedly in Nussbaum’s writings, but never explicitly theorised
 37

as I argue in the following chapters (2 and 3), is to miss the distinct and
substantial, if less orthodox, philosophical contributions of Sen’s own
version.
In this respect it is important to note the very significant differences
between Nussbaum’s project and Sen’s.39 Nussbaum is concerned to
produce a philosophically coherent (partial) theory of justice; Sen is
concerned with producing a general framework for evaluating the
quality of lives people can lead that can incorporate the very diverse
concerns and dimensions that may be applicable. Their foundations are
different: Nussbaum’s theory is based on the concept of ‘dignity’ and
she often criticises Sen’s emphasis on ‘freedom’ (e.g. Nussbaum 2003,
44–46). Nussbaum pays less attention to feasibility considerations,
instrumental linkages, or the independent significance of well-being,
being somewhat wary of “Utilitarian associations” (Nussbaum 2000, 11–
15). Despite the sufficientarian threshold, her approach is also
somewhat utopian in seeking full implementation since what she
considers minimal justice is specified so demandingly that no country
yet meets it (though she has suggested that Finland may be close). Sen’s
capability approach in its normative, ‘developmental’ aspect, is mainly
concerned with identifying and promoting practical comparative
improvements. It is a mistake to think that Nussbaum’s theory can do
the job of Sen’s capability approach better – it tries to do quite a
different job.

V. CONCLUSION
This chapter has discussed the development, structure and applications
of Sen’s capability approach, and outlined its take up by academic
philosophers. As Ingrid Robeyns, a former doctoral student of Sen,
relates, Sen sees the capability approach as operating at three distinct
levels (Robeyns 2000, 3). In order of importance these are:

or connected with the capability requirements of justice. Indeed, if we were to take it

seriously as doing real work in her account, then the character of her theory would
seem to be what Rawls terms a ‘comprehensive doctrine’, based on a metaphysical
conception of the good, rather than a political conception of the right, and thus in
tension with her current presentation of it as suitable for an “overlapping consensus”
(Crocker 2009, 188–92).
39
For an extended critical comparison see Des Gasper (Gasper 1997) and David Crocker
(Crocker 2009).
38

1. As a framework of thought;
2. As a critique on other approaches to welfare evaluation;
3. As a formula to make interpersonal comparisons of welfare.

It is as a framework of thought that the capability approach makes

its greatest contribution, and this will be the main focus of the
remainder of this dissertation. How we see the world matters for how
we can think about changing it for the better. Sen’s claim that in making
judgements of individual advantage and social welfare, “The basic
concern is with our capability to lead the kinds of lives we have reason
to value” (Sen 1999a, 285) is an invitation to understand well-being and
development in a new and better way.
The following two chapters elaborate on Sen’s methodology of
evaluation, on what kind of “framework” his capability approach
provides. The second half of the thesis considers what this framework
can achieve by examining and addressing three challenges to the
capability approach with respect to adaptation (4), development ethics
(5), and social justice (6).
 39

Chapter 2: Two Critiques of Sen

One of the distinctive features and, I think, strengths of Sen’s practical

philosophy in general is his commitment to a particular mode of
incomplete, open, and situated evaluation – namely ‘judgement’. Yet
many readers of Sen’s work (including both critics and admirers) appear
to understand this constitutive feature as a problem to be overcome.
They express great concern about the operationalisability of the
capability approach: that it is not in an appropriate form to be put into
application, either in theory or policy practice (e.g. Sugden 1993, 1553;
Rawls 2003a, 13 fn. 3). Operationalisation is an unfortunately ugly term
usually reserved for policy oriented discussions, and it is certainly true
that some development economists remain profoundly sceptical of the
capability approach because they either do not see what the real world
referents of concepts like capabilities and functionings are, or do not
see their relevance (e.g. T. N. Srinivasan 1994; Dasgupta 2005). However,
in an analogous sense, concerns about ‘operationalization’ are also a
central theme of the philosophical literature about Sen’s capability
approach: many critics say that it is too under-theorised to be put into
philosophical practice, while many sympathisers say it is too under-
specified (which comes to the same thing).
What I call the ‘operationalist approach’ has a dominant position in
the philosophical literature about the capability approach. It can be
characterised in terms of two features: theoretical reading and
theoretical reconstruction. The first, and most essential, is considering a
philosophical account in terms of a conventional theory. Of course,
‘theory’ can have various meanings, but the most relevant one here is “a
system of rules, procedures, and assumptions used to produce a result”
(World English Dictionary). The capability approach as Sen presents it is
manifestly unable to produce such results - definitive conclusions -
either at the theoretical level (which capabilities matter?; what would a
capability theory of justice look like?) or at the practical level (how to
treat particular cases).
Some operationalists discard the capability approach immediately on
establishing this (e.g. Roemer 1998, 191–3).40 Others will continue to
engage with it by reconstructing it into a theoretical form that they can
40
As John Roemer puts it, "One feels that Sen often tries to make a virtue of necessity
when he writes that certain hard questions have no right answers. The alternative, in
science, is to admit that there are answers, but we do not have them yet." (Roemer
1998, 193)
40

properly assess. So, in the second phase Sen’s account is reconstructed

to be more theory-like. The reconstructions are claimed to be
improvements on the original because they have a more orthodox
‘philosophical’ form, with clearly defined concepts whose operations are
defined by a theoretical structure. These reconstructions are then
developed (by proponents, such as Martha Nussbaum41) or contested (by
critics, such as Thomas Pogge) in standard analytical philosophy ways:
testing and elaborating their theoretical coherence, implications, and
relationship to other theories.
My concern is that this approach is driven by a prior and
unexamined commitment to a particular understanding of
operationalization, as asking what Sen’s capability approach can
contribute to the kinds of debates that moral and political philosophers
are already familiar with.42 Of course operationalisation is not in itself
an illegitimate way to approach Sen’s work. Its concerns and methods
reflect a strong and important tradition in analytical moral philosophy,
and its aim to try to understand what the capability approach is saying
in familiar terms is certainly legitimate. It can also be productive.
Negatively, it can show how and at what points Sen’s capability
approach deviates from orthodox methodology. It can show how
conventional readings of its core concepts have been tried and what
problems were found. It can point out lacunas in what the capability
approach addresses – aspects which fall outside what Sen’s account can
do. Positively, it can develop particular insights from Sen’s work in a
more orthodox way (as I showed in section IV of the previous chapter).
Yet I believe that the operationalization approach applied is
sometimes applied in an uncharitable way, producing distorted readings
of primary texts and ungrounded claims that are distracting rather than

41
“[T]he reader who looks for a fully formulated account of social justice generally,
and gender justice in particular, in Sen’s work will not find one; she will need to
extrapolate one from the suggestive materials Sen provides (Nussbaum 2003, 34).”
42
For example, as Ingrid Robeyns notes, “the literature on the capability approach does
not have a normative account of the distinction between personal and collective
responsibility, which is a core aspect of many theories of justice and social and public
policies. The absence of such an account of responsibility also limits the comparability
of the capability approach with theories of fair compensation in welfare economics or
responsibility-sensitive accounts of justice in political philosophy.” (Robeyns 2005,
192) Yet while the luck egalitarian approach may presently be dominant in egalitarian
theorising about justice it is not at all obvious why not contributing to its conversation
should be seen as a problem. As Elizabeth Anderson has argued quite forcefully, a
focus on what she calls “cosmic injustice” is hardly a requirement for egalitarian ethics
and can in fact be criticised as mistaken (Anderson 2010).
 41

helpful.43 The result is that the philosophical conversation about Sen’s

work does not always focus as well as it could on identifying his
particular contributions and the fair assessment, in its own terms, of the
capability approach’s qualities, potential, and scope for improvement.
For example it has generated a debate between perfectionist and liberal
interpreters of Sen’s capability approach that misreads Sen in significant
ways.
On the one hand, perfectionists such as Martha Nussbaum and
Séverine Deneulin have argued that in statements such as “The basic
concern is with our capability to lead the kinds of lives we have reason
to value (Sen 1999a, 285),” Sen is clearly referring to a positive vision of
the human good which must be more completely specified to be useful
(see for example (Nussbaum 1988; Nussbaum 1993; Deneulin 2002;
Arneson 2010)).44 Such writers question what guidance the capability
approach can give without an explicit description of what kind of life
“we have reason to value” (or at least a ‘roadmap’ of how to determine
it).
On the other hand, many political liberals from John Rawls to Robert
Sugden have complained that Sen’s emphasis on access to lives we have
reason to value implies a commitment to some ‘objective’ index for
evaluating advantage whatever those concerned might actually value
(Sugden 2006; Dowding 2006; Rawls 2005, 182–186; Dworkin 2002, 301–
2).45 In terms of Rawlsian political liberalism, Sen’s capability approach
is taken to imply the political endorsement of a particular conception of
the good whatever dissenting individuals believe, which fails the

43
This is of course a general problem of intellectual life. Adam Smith’s actual writings
for example have long been reduced to a few decontextualised anecdotes (butchers
and bakers, and the invisible hand) and distorted further by reading these only as
contributions to contemporary economic theory (cf Sen 2010a; Wells 2013 § 1; and, for
a case study, Schumacher 2012). Nor is it as uncommon as one might expect for this to
happen to writers in their own lifetimes, for example to Ronald Coase, whose work was
transformed, by others, into the theorem named after him (Yalcintas 2009).
44
It is true that Nussbaum now presents her account somewhat vehemently as political
rather than (Aristotelian) perfectionist, and has discussed the distinction at length
(Nussbaum 2003, 50; Nussbaum 2011a; Nussbaum 2011b, 6 fn 9). But I am not alone in
considering that the addition of Rawlsian devices like the ‘overlapping consensus’ has
not altered the substance of her account or its problems with justification (Claassen
and Düwell 2012). Deneulin defines perfectionism as “a moral theory that regards
certain activities, such as knowledge, health or artistic creation as good, independent
of any subjectivity” (Deneulin 2002, 498–9). I think this still describes how Nussbaum
justifies her list.
45
As John Roemer puts his challenge to Sen, "Who are you, Justice Commissar, to say
the Bengali beggar's capability is less rich than the Princeton professor's?” (Roemer
1998, 193)
42

requirement that “the public conception of justice is to be political, not

metaphysical” (Rawls 1985, 223).46
On the face of it this appears to present a material dilemma for Sen’s
capability approach: either the capability approach is politically liberal
or it is perfectionist. If it is politically liberal, it must respect individuals’
sovereignty to decide for themselves what the good life consists in, so it
should stop trying to evaluate advantage and instead support a
procedural approach of removing unfreedoms and providing general
purpose freedoms. If it is perfectionist, then it should specify and
justify its theory of value. From the perspective of analytic moral
philosophy, Sen’s refusal to go along with this vision of a stark choice
can be seen as weakness and ambiguity, rather than positively as
responsibility and flexibility.47 Although Sen has repeatedly declared his
lack of interest in building a coherent or complete theory of justice
(most recently in Sen 2009a), successfully dispelling such challenges
would seem to require something more: a worked out account of what it
is that Sen is trying to do, if not to provide a standard theoretical
account, and an analysis of its philosophical status.
The following chapter is concerned with providing such a positive
account of Sen’s system of evaluation as judgement, which I analyse
with the help of the account of judgement developed by Samuel
Fleischacker in A Third Concept of Liberty (1999). This chapter’s purpose
however is more negative, in the nature of brush-clearing. I will show
how the operationalization approach can systematically distort Sen’s
concerns, and even, sometimes, what he says. I will do so by engaging in
a close reading of critiques of Sen’s ‘under-theorisation’ by two
influential scholars. The first is by the Rawlsian justice theorist, Thomas
Pogge, who criticises the capability approach from outside (Section I).
The second is by Martha Nussbaum who, as a well-known exponent of
the capability approach, can be seen as an internal critic (Section II).

46
Of course, there is something of a contradiction in these critiques, since the
perfectionists see the account as underspecified and relying too much on a
presumption of individual agency, while the liberals accuse it of being overspecified in
a way that prevents individuals from exercising autonomy in choosing the good life.
47
Nussbaum, for example, has repeatedly challenged Sen to say whether he is a
“comprehensive” or “political” liberal (Nussbaum 2011a).
 43

I. POGGE’S CRITIQUE
Thomas Pogge is a political philosopher working within the Rawlsian
tradition who has long been engaged in research and activism in
support of global justice for the poor. He is a noted and vehement critic
of the capability approach (which he sees as having been developed
jointly by Sen and Nussbaum).48 Pogge notes that the increasing
influence of the capability approach has come at the expense of its main
alternatives, Rawlsian resourcism and utilitarianism, and sets out to
show why the choice of the capability approach cannot be justified. It is
interesting to see how he sets up the framework for his critical
assessment, in terms of satisfying a requirement for a “public criterion
of justice”.

Instead of asking which approach is superior, we should ask which

approach can deliver the most plausible public criterion of social
justice...Neither Sen nor Nussbaum has so far shown that the
capability approach can produce a public criterion of social justice
that would be a viable competitor to the more prominent resourcist
views. (Pogge 2010, 18)

Later the meaning of this becomes clear. Pogge means that any
competitor to Rawls’ account must have the same institutional focus
and theoretical scope and ambition as Rawls’ Theory of Justice.

We are seeking a public criterion of social justice that tells us how

an institutional order ought to be designed, and also how existing
institutional schemes fall short and how they should be reformed.
For this purpose we need not merely a partial ordinal ranking, but a
complete interval ranking. We need to know what positive or
negative resource compensation each participant should be entitled
to on the basis of his or her specific natural endowments. As an
institutional order is fully specific, so is the public criterion of
justice underlying it. Of course, Sen may reasonably believe that
there is a plurality of permissible public criteria of social justice
exemplifying the capability approach. But, for all Sen has published
on this topic, he has done little toward ruling out any candidates
within the vast space of conceivable capability views. So far, what he
has mainly proposed is a new language. (Pogge 2010, 51)

48
Pogge wrote a lengthy critique Can the Capability Approach be Justified (Pogge 2002),
and recently republished it in a slightly adapted and shortened version (Pogge 2010).
44

Pogge thus appears to be concerned with the capability approach

only at the level of theoretical competition with Rawlsian theory. His
understanding of the capability approach appears to derive from
toggling between a theoretical interpretation of Sen’s writing and more
explicitly theoretical material by Anderson and Nussbaum (to try to fill
in Sen’s theoretical gaps). This theory-driven perspective leads Pogge to
read and assess the capability approach literature in terms of an
instruction manual for designing a perfectly just institutional scheme, a
test which it of course fails: the extent to which the capability approach
provides a “new language” is the extent to which it fails to provide a
good - i.e. complete and rigorous - theory. As Sen points out in his reply,
Pogge’s terms of assessment and conclusions are driven not only by a
concern with theoretical qualities but by a particular idea of what a
theory of social justice should look like, which the capability approach
was not designed to meet.

Pogge assumes that the transcendental approach to justice is the

only way of having "a public criterion of social justice," and thus
focuses on "how an institutional order ought to be designed? (p.51).
This confined interest only in "the perfectly just," rather than in
public criteria for enhancing justice in the world in which we live,
takes Pogge to the conclusion that what is needed is "not merely a
partial ordinal ranking, but a complete interval ranking." (Sen 2010b,
250)

Pogge’s theoretical focus does bring up some interesting points. For

example he points out that the capability approach says nothing about
the organisation of the economy in which it would operate, either in
terms of considering the problem of scarcity of resources for capability
enhancement (Pogge 2010, 52), or the fair division of the burdens of
economic cooperation (Pogge 2010, 44). In the first part of the critique
he also argues for a high degree of theoretical convergence between the
resourcist and capability approaches.49 He argues that of course any
good resourcist should take into account factors like climate, that affect
people’s ability to convert resources into functionings, as part of the
calculation of how much resources “broadly conceived” people actually
have.50 As a result, Pogge is able to quickly dismiss the political

49
For an account of how a Rawls-Sen convergence might look from the other direction,
see Robeyns (Robeyns 2009).
50
Pogge is so successful in this that Ilse Oosterlaken argues that he is actually “a
capability theorist in disguise” (Oosterlaken 2012).
 45

liberalism critique (discussed above) as misconceived. Pace Rawls: “a

capability view can be formulated quite generally so that it focuses on
capabilities that (like Rawls' social primary goods) are important to all
human pursuits or nearly all” (Pogge 2010, 19–20).
Nevertheless, Pogge’s theory-driven (and at times frankly hostile)
reading of the capability approach also leads to some striking
distortions. Pogge argues that the real distinction between the capability
and resourcist approaches, and the reason we should opt for the
resourcist approach, is that resourcists are committed to specifying an
individual’s fair share in terms of “some conception of the standard
needs and endowments of human beings” rather than, as capability
theorists do, “the specific needs and endowments of each particular
person” (Pogge 2010, 23–4, original emphases). Here, Pogge’s
theoretical framing does serious harm to his reading of the capability
approach literature. It leads him to characterise the capability approach
as primarily concerned with the identification and compensation of
“naturally disfavored” individuals (e.g. Pogge 2010, 51). Neither of these
is correct, but both seem to follow from Pogge’s concern with analysing
the capability approach within the same theoretical framework as
(Rawlsian) resourcism.
First, Pogge characterises the capability approach as a search for
victims, for people who have such severe physical or mental
impairments that a standard helping of resources is insufficient for
them to achieve a minimally decent quality of life. For example, women
(Pogge 2010, 24). In doing so, Pogge appears to forget all he has already
discussed and accepted about how the physical and social environment
matters for translating resources into functionings (Oosterlaken 2012).
Sen, by contrast, argues consistently that the capability approach is
concerned with what people can do with what they have, which is quite
a different evaluative space.
Capability evaluation does not, pace Pogge, stigmatise certain people
as “naturally disfavored” by grading them in terms of the deficiencies of
their endowments.51 If it is used for comparing individuals it is in terms
of their capability, their access to lives they have reason to value, rather
than particular determinants of that. In terms of normative evaluation,
what the capability approach says is that if we take literacy, for example,

51
Curiously, Pogge also criticises the capability approach for failing to make this
stigmatising grading system sufficiently comprehensive to be workable (Pogge 2010,
51).
46

as equally important for all, then illiteracy becomes a morally significant

problem. But the causes of illiteracy - whether a child’s dyslexia (natural
endowment) or indifference (a moral failing?) or teacher absenteeism
(‘resources’) – are discriminated in empirical rather than normative
terms i.e. that they operate differently and would need addressing in
different ways. The capability approach lacks the nuanced moral
structure - Pogge seems to presume some form of luck egalitarian
welfare theory - to be able to make the invidious distinctions it is
accused of.
The mis-reading follows from Pogge’s retention of the Rawlsian
starting point of human beings “standard endowments” and
reconstruction of the capability approach as concerned with reducing
the deviation from that moralised benchmark, which he terms “vertical
inequality” (Pogge 2010, 44–48). In other words, personal heterogeneity
is merely an add-on to a theory of justice (and, according to Pogge, an
unnecessary one). Yet it seems unlikely that this is what Sen had in
mind, since he has directly criticised the standard endowments
assumption of the Rawlsian social contract approach on numerous
occasions.52
Second, Pogge repeatedly invokes a compensation principle which
implies seeing capability as an aggregate index on which everyone would
be scored and ranked, with additional resources provided in
compensation to those falling below some threshold. Pogge then moves
on to accuse the capability approach of promoting an ethic of
competitive victimhood, in which individuals strive to show that they
are the most helpless and pathetic and therefore most deserving of
compensation (Pogge 2010, 46). Pogge goes so far as to mock the
foolishness of taxing the “able-bodied” to provide compensation to
someone such as Stephen Hawking, who though severely disabled seems
to lead a rather more successful life than most.
However, Pogge’s compensation principle is based on his theoretical
reconstruction of the capability approach as concerned with aggregate
vertical inequality. In fact, multi-dimensionality is a constitutive feature

52
“Investigations of equality—theoretical as well as practical—that proceed with the
assumption of antecedent uniformity (including the presumption that 'all men are
created equal’) …. miss out on a major aspect of the problem. Human diversity is no
secondary complication (to be ignored, or to be introduced ‘later on'); it is a
fundamental aspect of our interest in equality.” (Sen 1992, xi)
 47

of the capability approach and appears everywhere in Sen’s work.53 The

capability approach is all about disaggregated analysis, of looking to see
what kinds of valuable functionings people do and do not have access
to. Thus the millionaire who cannot read in a literate society is lacking
something that he has reason to value, even if ‘all things considered’ his
life is a successful one that many would like to lead. The capability
approach does not identify the millionaire as a victim, but as someone
who is missing out on a dimension of the flourishing life. Because the
missing functioning is remediable it can be directly targeted (though
whether or not the millionaire should pay for his own reading lessons or
a personal assistant to do his reading for him is, as Pogge would be right
to point out, not addressed by Sen’s capability approach).
In any case, the capability approach is not a proposal for giving the
disabled an envelope of cash to bring their aggregate ‘capability-welfare’
over a threshold. And when one considers the Hawking case from the
capability approach it is clear that it is the remediable aspects of his
disability that have been targeted (he has a number of full-time
professional carers, not to mention his famous voice synthesiser and
wheel-chair) which allow him access, if not to specific functionings like
walking or talking, at least to the fundamental capabilities of mobility
and communication.
Pogge’s institutionalist-theory perspective also leads to quite
peculiar specific claims. Perhaps most egregiously, Pogge claimed that
the Human Development Index (HDI),54 which combines statistics
representing GNP per capita, longevity, and adult literacy to provide an
international ranking of countries’ development,

provides an incentive to concentrate resources on the healthier

individuals in such groups, whose life expectancy can be greatly
extended by ensuring only that they have the most basic nutrition,
sanitation, and medical care. The HDI thus encourages policy makers
to withhold scarce resources from those who have special needs that
make their life expectancy more expensive to extend. (Pogge 2002,
213)55

53
“To insist that there should be only one homogeneous magnitude that we value is to
reduce drastically the range of our evaluative reasoning.” (Sen 1999a, 77. See also 93-4)
54
The HDI was developed for the United Nations Development Programme by Mahbub
ul Haq in collaboration with Amartya Sen. Since 1990 it has been published in the
annual United Nations Human Development Reports, which are closely associated with
Sen’s capability approach to human development.
55
Pogge did not repeat his critique of the HDI in the 2010 version of his paper. Several
of his specific technical concerns with the statistical construction of the index seem to
48

It seems that Pogge interprets the HDI as the capability approach in

action, and considers that it will lead ranking-focused governments to
preferentially transfer resources to ‘capability wizards’ who are more
efficient at transforming resources into longevity. Pogge here makes
rather grand assumptions about both the power and role of the HDI. Do
governments actually have an “incentive” to do what Pogge suggests?
Pogge provides no evidence that they do. Is Pogge correct that this how
the HDI is intended to be used? Not according to those who have
worked on it, who consider it to be a tool for evaluation rather than
prescription. As Jeni Klugman (the director of the Human Development
Report Office) and co-authors put it,

an index of capabilities is conceptually different from a social

welfare function. The key difference is that a social welfare function
is designed to be maximized, and thus the trade-offs along that
social welfare function can be interpreted as values. But a
capabilities index is meant to give a measure of the extent to which
people in different countries have access to substantively different
lives. Obviously, we care about the expansion of these capabilities
and believe that expansion is welcome. But this is very different
from claiming that the maximization of capabilities should be the
only objective of social action….. In this sense, the capabilities
approach contrasts with traditional theories of social justice, such as
utilitarianism, which postulate an overarching objective as the end
result of human action. The capabilities approach is a partial theory
of well-being that does not purport to establish a complete
description of all the components of a good life…. (Klugman,
Rodríguez, and Choi 2011, 29–30 original emphases)

The HDI was conceived as an aid to evaluation that makes available

more (though still quite basic and limited) information about well-being
in many countries than previous international comparisons, which
typically focussed narrowly on per capita average income (Sen 2004a,
79). Sen himself has written about this purpose and the limitations of
the HDI in many places and in more or less technical depth. So Pogge’s
charge is really rather puzzling. Sen notes, for example, that since the
HDI was specifically created to compete for public attention with income
per capita rankings - such as published in the World Bank’s World
Development Reports - it is not surprising that it is a crude and
imperfect measure of quality of life (Sen 1999a, 103 fn. 41). The

have been at least somewhat addressed in the 2010 revamp of the index (Klugman,
Rodríguez, and Choi 2011).
 49

appropriate question is, is it a better measure than what came before it?
Sen also points out that the simplistic aggregative index is published as
only one part of an annual Human Development Report which includes
much wider ranging empirical and qualitative analysis (ibid).
Given the very wide availability of such information about the HDI’s
evaluative orientation, it seems to me that Pogge’s critique best makes
sense as following from, rather than informing, his institutional-
theoretical approach. On Pogges’s reading the HDI is a social welfare
function that should be maximised because it is a metric, and, in the
kind of institutionalist theoretical framework that Pogge takes for
granted, metrics are targets for an institutional scheme to fulfil. They
must therefore be precisely focussed on what the correct theory of
justice considers relevant, and their implications well mapped, or they
will have disastrous consequences.
I have so far said a great deal about how I think Pogge has misread
the capability approach. However, it is worth noting that Pogge does in
fact pick out features of the capability approach that I think are central
to its proper understanding (Pogge 2010, 50–1). He just doesn’t think
they are important, because they don’t qualify as material for a
competing theory to resourcism. Pogge suggests that the capability
approach can play two roles with respect to a proper (resourcist) theory
of justice: evidentiary (which he associates most particularly with Sen’s
work) and heuristic (Nussbaum’s list). The evidentiary role concerns the
identification of capability shortfalls and their causal tracing to identify
mistakes in how fair resource bundles were calculated or sources of
injustice in need of removal. The related heuristic role is that it can help
us think through what the content of a theory of justice should be
concerned to achieve:

It can help us think of all the personal and public goods and
supports that human beings need to flourish fully, from the school
curriculum to the organization of workplaces and organs of
democratic decision-making....not as the metric within a public
criterion of social justice, but as a useful guide in the development
of such a criterion (Pogge 2010, 50).

I will return to these positive features in the following chapter. For

now let us turn to an internal critic of Sen’s account.
50

II. NUSSBAUM’S CRITIQUE

As I noted above, a number of philosophers attracted (‘recruited’) to
Sen’s capability approach, particularly by its forceful rejection of
utilitarian and resourcist approaches to the ‘equality of what’ question,
nevertheless consider Sen’s account too underspecified to make a real
contribution. Firstly, its positive account is not worked out sufficiently
for the demands of normative theorising, translating moral intuitions
into definitive propositions with clear implications. Secondly, as the
saying goes, ‘It takes a theory to beat a theory’. In order to make
headway in the academic philosophical competition against more
established highly theorised accounts, or even to be taken seriously as a
competitor, the capability approach needs to meet certain standards of
theoretical consistency, coherence, and so forth.
Nussbaum, has taken up this challenge full-heartedly, and her
(partial) capability theory of justice attempts to operationalise Sen’s
capability approach in exactly this way. Here, for example is how she
put the challenge in her first foray into the capability approach.

Getting the list of functionings that are constitutive of good living is

a matter of asking ourselves what is most important, what is an
essential part of any life that is going to be rich enough to count as
truly human....Sen needs to be more radical than he has been so far
in his criticism of utilitarian accounts of well-being, by introducing
an objective normative account of human functioning and describing
a procedure of objective evaluation by which functionings can be
assessed for their contribution to the good human life. (Nussbaum
1988, 175–6)

For Nussbaum, critical evaluation is not enough to support a theory

of justice that will provide the necessary rigorous structure to provide
the compelling arguments and direction to get things done. Which
capabilities are most significant and should be political goals?

[T]here is no way to take the capabilities approach forward, making

it really productive for political thought about basic social justice,
without facing this question and giving it the best answer one can.
(Nussbaum 2003, 50)

Nussbaum’s critical engagement with Sen’s capability approach is

motivated both by a demand that it produce definitive conclusions
about principles, and a demand for justification: the idea that if the
 51

capability approach is evaluating something important, it should be

possible to identify and justify that independently of the evaluation
process. The particular theoretical moves Nussbaum makes align closely
with her theoretical background in Aristotelian philosophy and the
philosophy of (constitutional) law. The metric and rule of her capability
theory of justice are universal access to at least minimal levels of the 10
central capabilities she identifies with the requirements of human
dignity, while its implementation takes a legal institutional
(constitutional) form.56
Nussbaum’s capability theory of justice has received quite intense
criticism. Some have questioned the epistemological basis of her
approach, suggesting that her list remains too Aristotelian (Clark 2002);
and that Nussbaum’s claims about the universality of the list items
being supported by “years of cross-cultural discussion” are actually
quite thin (Stewart 2001; Okin 2003). Others have argued that her legal-
moral-philosophical orientation is elitist and over-optimistic about what
constitutions and governments are like and are capable of (Menon
2002); is over-specified and paternalistic, eclipsing cultural values
(Deveaux 2002); yet still misses out important capabilities (Robeyns
2005, 207), and is inappropriate for many uses, such as quality of life
measurement (Robeyns 2005) or development fieldwork (Alkire 2005,
35–45).
Some of these criticisms are important, but much of their
motivation, and the emotional vehemence with which they are given,
relate to a perception that Nussbaum is imposing a definitive capability
theory on everyone.57 This may be due to Nussbaum’s full-hearted and
self-confident promotion of her theory – she clearly believes that it is
both substantively correct and the best way to take the capability
approach forward. However, she has also emphasised that she is not
trying to impose a definitive capability theory on everyone. In nearly
every exposition of her account she makes a clear and explicit
distinction between the dimensions of justification (why her theory is

56
See the earlier discussion, in chapter 1:IV.
57
Hence, “What is deeply puzzling is Nussbaum’s calm assumption that she has
arrived at the list of good things no good person could possibly not want for everyone”
(Menon 2002, 157 original emphases).
52

best) and implementation (its more humble meta-status as an object for

democratic deliberation and decision by those concerned).58

I insist on separating issues of justification and implementation. I

believe that we can justify this list as a good basis for political
principles all round the world. But this does not mean that we
thereby license intervention with the affairs of a state that does not
recognize them. It is a basis for persuasion.... (Nussbaum 2004, 198)

I think that the core issue is not the strength of Nussbaum’s

theoretical contributions, which as Pogge correctly points out are
heuristically helpful in thinking about what the flourishing human life
might require.59 Rather, the problem is the status of her theory (or any
other) as a candidate to be the correct theory of capability, rather than
being more or less helpful in certain contexts and for certain purposes.
Nussbaum’s approach presumes that without theoretical specification
the capability approach cannot be operationalised, and therefore she is
led to claim that “Sen cannot avoid committing himself to a core list of
fundamental capabilities [such as this one]” (Nussbaum 2001, 46). Her
2003 paper on Capabilities as Fundamental Entitlements provides a
particularly clear illustration of this which is worth a close reading.60 The
abstract reads in full,

Amartya Sen has made a major contribution to the theory of social

justice, and of gender justice, by arguing that capabilities are the
relevant space of comparison when justice-related issues are
considered. This article supports Sen’s idea, arguing that capabilities

58
In Women and Human Development she addresses this issue at some length
(Nussbaum 2000, 101–5). However, it remains fair to say that Nussbaum’s vision of
how implementation would proceed is rather lightly sketched.
59
An additional contribution, from her first publication on the capability approach
(Nussbaum 1988), is an interesting distinction between three different aspects of
capability as ‘basic’ (naturally endowed), ‘intrinsic’ (capacities attained in the life of an
individual) and ‘combined’ (when external circumstances make their exercise possible)
that would seem generally helpful for capability analysis.
60
In her recently published textbook on the capability approach, Creating Capabilities
(Nussbaum 2011a), Nussbaum takes a more conciliatory line with respect to Sen’s
account. She presents the capability approach as having a complementary division of
labour between the comparative assessment of quality of life (which she considers
Sen’s principal concern) and social-justice theorising (her own concern). However, I do
not believe this represents a real shift in Nussbaum’s views since the 2003 paper,
because she goes on to criticise Sen’s approach in the same way (often verbatim). I
have nevertheless used the older work for this exercise because this recent book seems
to have been written quite quickly and did not seem a suitable candidate for close
exegesis.
 53

supply guidance superior to that of utility and resources (the view’s

familiar opponents), but also to that of the social contract tradition
and at least some accounts of human rights. But I argue that
capabilities can help us to construct a normative conception of
social justice, with critical potential for gender issues, only if we
specify a definite set of capabilities as the most important ones to
protect. Sen’s “perspective of freedom” is too vague. Some freedoms
limit others; some freedoms are important, some trivial, some good,
and some positively bad. Before the approach can offer a valuable
normative gender perspective, we must make commitments about
substance. (Nussbaum 2003, 33)

In the remainder of this section I will follow up two of the criticisms

of Sen that Nussbaum raises in this paper, one major and one minor.
The first (major) issue concerns her characterisation of the role of
theory in thinking about social justice; the second (minor) issue is a
specific mis-reading of Sen’s understanding of freedom that I include to
illustrate my point that the same distorting effects of theory-driven
interpretations we saw with Pogge can afflict even readers highly
sympathetic to Sen’s concerns.
First, Nussbaum argues that in order to criticise aspects of the world
and social arrangements as unjust, we first need to have a worked out
normative theory.

[I]f the issue of social justice is important, then the content of a

conception of justice is important. Social justice has always been a
profoundly normative concept, and its role is typically critical: we
work out an account of what is just, and we then use it to find reality
deficient in various ways. Sen’s whole career has been devoted to
developing norms of justice in exactly this way, and holding them up
against reality to produce valuable criticisms. It seems to me that his
commitment to normative thinking about justice requires the
endorsement of some definite content. One cannot say, “I’m for
justice, but any conception of justice anyone comes up with is all
right with me.” Moreover, Sen, of course, does not say that. He is a
radical thinker, who has taken a definite stand on many matters,
including matters of sex equality. He has never been afraid to be
definite when misogyny is afoot, or to supply a quite definite
account of why many societies are defective. (Nussbaum 2003, 47,
emphases added)

This claim seems both logically and descriptively inadequate. It is

logically inadequate because it asserts that without having a fully
specified account of justice one cannot criticise injustice, and yet
54

accepts that Sen has apparently been going around doing just that for
some time. It is descriptively inadequate in its assertion that “Sen’s
whole career has been devoted to developing norms of justice in exactly
this way”. In fact Sen has rejected this kind of ‘theory first’ requirement
for thinking about justice explicitly and in many places61 (a point I will
not elaborate on further here as it is the focus of the following chapter).
It seems to me that because Nussbaum already has this idea that we
need a substantive theory of justice to perceive injustice - strikingly
similar to Pogge’s perspective - she is led to reconstruct Sen’s project in
those terms, however strange the fit with what Sen actually says.
A second and more specific misreading is in Nussbaum’s critique of
Sen’s freedom terminology. Early in the paper, Nussbaum criticises the
limited “neo-liberal” concept of rights based on the key idea of “negative
liberty” understood in terms of the absence of state interference in an
individual’s life (Nussbaum 2003, 38–9). She argues that the capability
approach offers a superior understanding of rights “by focusing from
the start on what people are actually able to do and to be” (39). So it is
distinctly puzzling that a few pages later (44-6) she engages in a
blistering critique of Sen’s use of the language of freedom (particularly
in his Development as Freedom) to describe what the capability
approach is about.

Sen speaks throughout the work of “the perspective of freedom” and

uses language, again and again, suggesting that freedom is a general
all-purpose social good, and that capabilities are to be seen as
instances of this more general good of human freedom. Such a view
is not incompatible with ranking some freedoms ahead of others for
political purposes, of course. But it does seem to go in a problematic
direction. (Nussbaum 2003, 44)

I find this reading very strange, since Sen’s use of the word
‘freedom’ in Development as Freedom doesn’t seem to me to change the
meaning of the capability approach in any significant way (see also

61
For example in Development as Freedom (Sen 1999a, 254). Although of course, Sen’s
most systematic elaboration of his position has come relatively recently (Sen 2006b;
Sen 2009a): “[A] theory of justice that can serve as the basis of practical reasoning
must include ways of judging how to reduce injustice and advance justice, rather than
aiming only at the characterization of perfectly just societies – an exercise that is such
a dominant feature of many theories of justice in political philosophy today . . . The
assumption that this comparative exercise cannot be undertaken without identifying,
first, the demands of perfect justice, can be shown to be entirely incorrect.” (Sen
2009a, ix)
 55

Qizilbash 2005, 52–7). A person’s freedom (capability) is understood, as

before, in terms of their ability to “lead the kind of lives they have
reason to value” (Sen 1999a, 10). Further, rather than speaking of
freedom as “a general all purpose social good”, Sen repeatedly calls
attention to the unconventional multi-dimensional character of freedom
as he is employing it, and refers the reader to his more explicit technical
treatments of different aspects of this elsewhere (e.g. Sen 1999a, 290–
292).
It appears that Nussbaum is particularly sensitive to the word
‘freedom’ because of its connotations with particular ideologies (like
‘neo-liberalism’) or ethical theories (like libertarianism) of which she
strongly disapproves.62 Thus, as she chooses to define what Sen means
by freedom: “any particular freedom involves the idea of constraint: for
person P is only free to do action A if other people are constrained from
interfering with A (Nussbaum 2003, 44).” Nussbaum goes on to suggest
that Sen’s use of the word freedom creates new problems for the
capability approach: how to identify and treat i) the limits of freedoms
(“the freedom of businesses to pollute the environment [vs.] the
freedom of citizens to enjoy an unpolluted environment”); ii) which
freedoms are politically central (and which are not, like, “the freedom of
rich people to make large campaign contributions”); and iii) which
freedoms may be invidious (such as the “male freedom” to rape one’s
wife or sexually harass one’s female co-workers) (Nussbaum 2003, 44–5).
Now I think it should be fairly obvious that it is not the word
‘freedom’ that generates these ‘problems’ (which would apply equally, if
at all, to the word ‘capability’).63 Rather, this critique seems to be driven
by Nussbaum’s focus on the requirements of operationalization. Like
Pogge she worries that an underspecified capability theory of justice
would be open to abuse or useless:

62
Compare also with Nussbaum’s dismissal of Sen’s distinction between the well-being
and agency aspects of evaluation because of its unfortunate “Utilitarian associations”
(Nussbaum 2000, 14)
63
Indeed, Nussbaum’s anthropological justification of her own list in terms of the
requirements of a truly human life has a similar problem, since she would like to
exclude certain all too human functionings like cruelty and aggression (Claassen and
Düwell 2012, sec. 2). “Not all actual human abilities exert a moral claim, only the ones
that have been evaluated from an ethical viewpoint. (The capacity for cruelty, for
example, does not figure on the list.)” (Nussbaum 2000, 83). That is, Nussbaum’s own
approach requires a second ethical evaluation stage to filter the results of the
anthropological stage that generates candidates for morally significant functionings.
56

Either a society has a conception of basic justice or it does not. If it

has one, we have to know what its content is, and what
opportunities and liberties it takes to be fundamental entitlements
of all citizens. One cannot have a conception of social justice that
says, simply, “All citizens are entitled to freedom understood as
capability.” Besides being wrong and misleading in the ways I have
already argued, such a blanket endorsement of freedom/capability
as goal would be hopelessly vague. It would be impossible to say
whether the society in question was just or unjust. (Nussbaum 2003,
46–7)

To sum up, Nussbaum assumes a certain model for normative

theorising about justice which emphasises the requirements for what I
have called philosophical operationalization. The task of philosophers is
to develop theories which specify the requirements of justice
sufficiently to allow proper philosophical scrutiny and comparison with
other competing theories in order to determine which is the best theory
of justice. Only then is the theory to be applied to cases to say whether
or not they are just and, if not, what justice requires to be done. This, as
we saw above, is the same project, with the same rules, that Pogge is
engaged in. But it certainly is not Sen’s project.
What Sen is concerned with, and why he rejects the architectonic
role for theory promoted by Pogge and Nussbaum, is the subject for the
next chapter. I will argue there that Sen’s concern with coming to
judgements of individual advantage and social welfare, rather than
developing a capability based theory of justice, means that he relates to
theory in quite a different way. For him theories are better or worse
resources for grasping different aspects of complex and opaque ideas
like poverty and well-being. Since many different theoretical frames and
lists may be helpfully informative in different cases and for different
purposes (to examine the character of extreme poverty in developing
countries for example (Alkire and Santos 2010a), as opposed to gender
inequality in Western countries (Robeyns 2003)), this makes him both a
committed pluralist about theory selection and pragmatic in his use of
theory. Justice theorists, in contrast, are seeking the best all-round
theory of justice. This brings at least two features that don’t apply to
Sen’s project.
First, the structural requirements of a full theory (in terms of
coherence, consistency, and perhaps feasibility) leads justice theorists to
‘bite bullets’ by making decisions about which moral intuitions to focus
on and which to exclude. For example, in her capability theory of justice
 57

Elizabeth Anderson focuses on equality in the space of democratic

equality, and excludes “subjective” functionings from consideration in
order to meet the requirements for a theory of justice comparable with
that of Rawls (Anderson 2010, 84–6). Yet there is something odd about
being required to see someone who is desperately unhappy as doing
fine, merely because one’s theory rules out the consideration of such
information. It’s a little like wearing spectacles that only allow you to
see the colour blue. As Sen notes, “A state of affairs is informationally
rich. There is no particular reason to insist on an impoverished account
of a state of affairs in evaluating it (Sen 2000a, 491).”
Of course, the restrictions imposed by conventional theorising about
justice do present such a “particular reason” for deliberately narrowing
one’s vision. But Sen argues that one doesn’t have to accept the need for
such theorising in the first place. Because Sen is unburdened by the
constraints that come with conventional theorising about justice, he
doesn’t have to make definitive a priori decisions about excluding
certain kinds of information (not even that capability should be the only
dimension of evaluation) but can maintain an enviable openness about
the scope of evaluation.
Second, justice theorists are generally looking for the best answer to
the question of social justice, whereas Sen is only looking for helpful
viewing points. Justice theorists tend to think in terms of a contest
between whole theories, and to evaluate them in terms of their ability to
survive challenges to their coherence and how well they deal with
difficult cases. This focus on theoretical virtues can come at the expense
of substantive relevance (for example in debating such narrow issues of
principle as are involved in the esoteric examples of plovers’ eggs and
surfer bums). In contrast, since Sen is not engaged in building or
choosing the best single theory of justice, he can and does make
pragmatic use of a wide range of theoretical resources, from libertarian
accounts of the value of liberty to Marx’s idea of false consciousness,
without having to make a commitment to the whole theory from which
they come.64

64
Sen’s theoretical pragmatism is apparent across his work, for example in his
shockingly unorthodox but rather commonsensical combining of consequentialist and
deontological concerns under ‘comprehensive consequentialism’ as a contribution to
the 1970s meta-ethical debate between philosophers, and between philosophers and
economists (Thompson 2010).
58

III. CONCLUSION
This chapter has identified a particular orientation to operationalization
popular in mainstream analytical philosophy and in the literature on the
capability approach. I have argued that although this approach has its
merits, if used without care it can lead to a systematically distorted
representation of Sen’s project and thus misdirected assessments and
critiques. I illustrated this problem by closely examining two critiques of
Sen’s project, from the work of influential and well-regarded moral
philosophers noted for their engagement with Sen’s capability approach.
I showed that in the texts examined not only did Pogge and Nussbaum
tend to interpret Sen’s work in terms they were familiar with, but on
occasion (presumably without realising they were doing so) they also
appeared to do some violence to Sen’s own words in order to squeeze
the capability approach into the theoretical framework they expected.
While the cases I looked at were particularly striking, this phenomenon
is disappointingly widespread in the philosophical literature about the
capability approach.
Key to the operationalist approach is the theoretical reconstruction
of Sen’s supposedly obscure position into a philosophically orthodox
form. In the following chapter I will myself be attempting something of
this sort – while remaining acutely conscious of the difficulties in doing
so. After clearing the ground here and showing indirectly what the
capability approach is not, by showing how theory-driven
interpretations fall short, I will now turn to providing a positive account
of what I think is central to Sen’s methodology in the capability
approach and in other parts of his work: judgement.
 59

Chapter 3: Judgement in Sen’s Capability Approach

Rather than simple social facts waiting to be discovered, poverty and

development are complex and ill-structured issues, which cannot be
fully captured by the cognitive tools of a single discipline nor
subjected to standard methods. (St. Clair 2007, 148)

The previous chapter was concerned to demonstrate the flaws in one

popular approach to understanding Sen’s capability approach, in terms
of ‘operationalisation’. The aim of this chapter is to provide a positive
account of what it is that Sen does aim at: a distinctive style of critical
evaluation that I comprehend under the term ‘judgement’ in order to
place it within the existing philosophical literature. Specifically, I argue
that Sen’s approach to evaluation can be helpfully understood in terms
of the analysis of judgement developed by Samuel Fleischacker in A
Third Concept of Liberty, “a complex skill that draws on what we do in
aesthetic interpretation, in sorting through empirical evidence, in
making decisions in the common law, and in evaluating our ends for
cogency and value” (1999, 8).
This concept of judgement explains and justifies the apparent
oddities in Sen’s account that were the focus of the operationalist critics
just discussed, but it does so, I will argue, in a manner much more in
line with Sen’s stated aims, practice, and methodological reflections.
Recall that Sen not only refuses to provide a systematic normative
theory, or even merely a list, specifying what the capability approach is
concerned with. He quite definitively rejects the goal of constructing the
best capability theory or list as unnecessary, illegitimate, and counter-
productive (Sen 2004a). And yet Sen frequently makes definitive
statements about deprivation and injustice – for example, criticising
discrimination against women (e.g. Sen 1990d) or India’s failure to
provide literacy and nutrition to its children (Sen 2008b) - which would
appear to have substantive normative commitments behind them.
If one understands the centrality of judgement in Sen’s evaluative
project, one can see how these aspects are both consistent and
philosophically respectable. A judge is not concerned with choosing the
best overall theory, but with the best understanding of how to treat
particular cases in terms of the various theoretical resources at her
disposal. Judgements themselves have the key features of being explicit
in their reasoning and defeasible in their conclusions, and so are quite
different from subjective matters of opinion. I will argue that this is how
60

Sen’s methodology of evaluation proceeds. Appreciating the centrality of

judgement in Sen’s capability approach (and, indeed, his practical
philosophy in general) should focus attention on the requirements,
capacities, and character of such a judge. Sen has suggested that Adam
Smith’s impartial spectator provides a model for such a judge, most
extensively in The Idea of Justice, but he has only lightly sketched out its
role. Examining and elaborating on Sen’s use of the impartial spectator,
thus far relatively neglected in the literature on Sen’s work, thus seems
an important focus for better understanding the potentials and
limitations of Sen’s methodology of evaluation, and I take up that task
in the following chapter.
The argument of this chapter proceeds by bringing together Sen’s
writings on different areas which bear on this issue and taken together
reveal an underlying consistency. In the following section I outline in
more detail Sen’s ‘puzzling’ use of the capability approach. Section II
examines Sen’s explicit methodological reflections on evaluation.
Sections III and IV are brief case studies of how Sen’s distinctive
approach to evaluation plays out in two other spaces not directly
connected to the capability approach, how to think about justice (III) and
the entitlement approach to food security (IV). Section V introduces
Fleischacker’s schema of ethical judgement and uses it to organise this
material and also to show the philosophical legitimacy and strength of
Sen’s approach to evaluation. Section VI considers the implications of
this way of seeing Sen’s account of the capability approach.

I. SEN’S PUZZLING AMBIVALENCE

In Development as Freedom Amartya Sen presents a parable (Sen 1999a,
54–55). Annapurna has a choice between three different unemployed
labourers who would like the job of cleaning up her garden. Dinu is the
poorest, Bishanno is the most unhappy, and Rogini has a chronic
ailment that her wages would allow her to cure. Who would be the right
person to employ?
A purely resource-egalitarian, utilitarian, or capability theory of
justice would each focus on a different feature as the most ethically
salient and would straightforwardly indicate the corresponding person
as most deserving of the job. However, Sen doesn’t use the parable to
argue that a capability theory provides the right answer to the
 61

question.65 Rather, he uses it to demonstrate how the prior choice of a

theoretical perspective determines in advance which information about
the case is considered, and which is not.
In the constructive part of the chapter, Sen goes on to argue that
none of the main ethical accounts engages with a particular aspect of
human lives - “the freedom to achieve actual livings that one can have
reason to value” (Sen 1999a, 73 original emphases) - that would seem to
have prima facie significance. This aspect is of course the particular
concern of the capability approach “at the foundational level”, but it is
not its only concern in practice (Sen 1999a, 81). Sen is keen to stress
that he is not simply providing another closed theoretical account that
would evaluate cases on the basis of a narrow class of information.66

The capability perspective is inescapably pluralist. First, there are

different functionings, some more important than others. Second,
there is the issue of what weight to attach to substantive freedom
(the capability set) vis-a-vis the actual achievement (the chosen
functioning vector). Finally, since it is not claimed that the capability
perspective exhausts all relevant concerns for evaluative purposes
(we might, for example, attach importance to rules and procedures
and not just to freedoms and outcomes), there is the underlying
issue of how much weight should be placed on the capabilities,
compared with any other relevant consideration. (Sen 1999a, 76–7)

Sen’s capability approach is not just about content, but also method.
It is an approach to evaluation that is not only multi-dimensional (the
capacious capability space) but also multi-principled.67 This makes the
application of Sen’s capability approach particularly challenging, and, it
appears, quite deliberately so. As Sen explains, there is no “magic
formula” or “royal road” by which this weighting may be determined
65
Though for some reason many readers think that he does. Pogge for example claims
that “Sen invites us to conclude that Annapurna should give the job to Rogini”, and
then goes on to argue, based on further rather puzzling exegesis, that in any case the
argument of the parable is inconclusive “because it falls in the domain of what Sen
calls personal ethics rather than political philosophy” (Pogge 2002, 36). In The Idea of
Justice Sen returns repeatedly to a similar parable with respect to the question of who
should receive a flute, for similar purposes (Sen 2009a).
66
On this point, recall Robeyns’ identification of a hierarchy of goals for Sen “1. As a
framework of thought; 2. As a critique on other approaches to welfare evaluation; 3. As
a formula to make interpersonal comparisons of welfare” (Robeyns 2000, 3).
67
Hence, unlike its alternatives the capability approach “can take note of, inter alia,
utilitarianism's interest in human well-being, libertarianism's involvement with
processes of choice and the freedom to act and Rawlsian theory's focus on individual
liberty and on the resources needed for substantive freedoms (Sen 1999a, 86).”
62

technocratically without explicit public debate and deliberation “since

the issue of weighting is one of valuation and judgment, and not one of
some impersonal technology” (Sen 1999a, 79). So, unlike its competitors,
Sen’s capability approach does not dictate which of the three
unemployed people you should give the job to. Rather it gives you more
things to worry about since it suggests that decisions such as these
should be explicitly justified on the basis of how appropriately all the
information in a case is considered, rather than on the basis of
analysing what follows directly from an a priori choice of the right
(perfectly internally coherent) theoretical account of justice. As we will
see, this focus on open and objective evaluation is a general feature of
Sen’s methodology of evaluation.

II. SEN’S METHODOLOGICAL REFLECTIONS ABOUT EVALUATION

a. Open Evaluation

Recall that Sen’s criticism of alternative approaches to evaluating

advantage (such as utilitarianism or resourcism) was not only directed
at their content - that they were focused on the wrong things. That
problem is addressed by Sen’s promotion of the capability space as a
better focus for evaluation. However Sen also argued that these
approaches were methodologically flawed because they were
unnecessarily narrow in their informational basis. They confused a good
answer to the question, ‘Is this person well off in welfarist/resourcist
terms?’ with a good answer to the question, ‘Is this person well off?’.
That issue is addressed by Sen’s methodology of evaluation.
Sen has published several papers analysing the informational basis
for the application of moral principles (see e.g. Sen 1979d; Sen 1985a,
pt. I). As he puts it,

Informational analysis can be used to bring out the content, scope,

and limitations of different moral principles. Each moral principle
needs some types of information for its use, and -- no less
importantly -- "rules out" direct use of other types of information. In
their latter role, moral principles impose "informational constraints,"
demanding that certain types of information should not be allowed
to influence the moral judgments we make. (Sen 1985a, 169–170)

Sen has shown the significance of apparently reasonable a priori

restrictions on what kinds or dimensions of information to consider in
 63

an evaluation which committing to a particular theoretical approach in

advance usually entails. For example, ‘narrow consequentialists’ rule out
the elements of process, agents, and motivations in assessing the final
outcome, while ‘narrow deontologists’ rule out everything but
principles. To employ a phrase coined by Marcel Boumans and Mary
Morgan in the context of economic methodology (Boumans and Morgan
2001, 12–4), rather than a ceteris paribus clause stating that a claim will
hold ‘other things being equal’, an exclusion principle operates more as
a ceteris neglectis clause. It can be understood as a specific form of
ceteris paribus asserting that ‘other things are negligible’.68
A good illustration of this can be found in Sen’s systematic and
extended critique of what he has termed ‘welfarism’ in terms of the
narrowness of its informational base. Restricting evaluation only to
utility information means that one cannot distinguish between sources
of utility, for example between the cases of redistributive taxation and
sadistic torture if the utility numbers come out the same (Sen 1979a,
472–474). Sen’s famous paper, The Impossibility of a Paretian Liberal
(Sen 1970a), can be seen as an elegant demonstration of this limitation
of welfarism. He demonstrated there that even weak Paretianism based
on utility dominance (the mildest version of welfarism) is incompatible
with minimal liberty: that individuals be free to make certain decisions
in a private domain, such as what they should read. This “shows how a
variation of non-utility description can precipitate different moral
judgments even when the utility description is unaltered....contrary to
the essence of welfarism (Sen 1979a, 482).” Protecting the moral
principle of individual liberty - or other principles that are not
instrumentally related to welfarism, such as non-exploitation - requires
going outside the welfarist framework.
From this ‘informational perspective’, the question raised by
utilitarianism “is not whether well-being is an intrinsically important
variable for moral analysis, but whether it is uniquely so (Sen 1985a,
186).” In other words, moral theories must justify their exclusions as
well as their inclusions. Sen’s concern is that while such informational
restrictions allow theoretical consistency and elegance, and arguably
make answering questions easier (because one always gives the same

68
Hence one of Sen’s concerns about a canonical list of valuable capabilities: “To
decide that some capability will not figure in the list of relevant capabilities at all
amounts to putting a zero weight on that capability for every exercise, no matter what
the exercise is concerned with, and no matter what the social conditions are. This
could be very dogmatic.....” (Sen 2004a, 79)
64

kind of answer, no matter what the question), such a closed approach

fails in its responsibility to address the actual complexity that real world
cases present. This follows from the axiomatic approach itself which
seeks to reduce all problems to comprehension by an a priori account: a
particular theoretical account is endorsed in advance, and this places
artificial restrictions on the dimensions of information which are
deemed relevant and how they may be treated. As a result, Sen points
out, approaches which attempt to provide a consistent systematic rule-
based framework for dealing with all problems that may come up can
easily be shown to produce absurd or awful evaluations of plausible
examples.69
If the subject of an evaluation exercise is an issue of broad scope
and complexity, of how well people’s lives are going, for example, then
numerous aspects will be of prima facie relevance. Reasoned evaluation
requires maintaining openness to that informational richness for as long
as possible and justifying one’s choices of perspectives (informational
inclusions and exclusions) with respect to the demands of the case and
the purposes of the evaluative exercise. The evaluation of a complex or
ambiguous concept – like well-being or inequality - must respect its
underlying complexity rather than impose an artificial completeness and
precision that is inappropriate. Thus,

[I]f an underlying idea has an essential ambiguity, a precise

formulation of that idea must try to capture that ambiguity rather
than attempt to lose it. Even when precisely capturing an ambiguity
proves to be a difficult exercise, that is not an argument for
forgetting the complex nature of the concept and seeking a
spuriously narrow exactness. In social investigation and
measurement, it is undoubtedly more important to be vaguely right
than to be precisely wrong. (Sen 1989a, 45)

As Robert Sugden summarises Sen’s view, “The concept of a good

life is objective but fuzzy; there is no single, clearly correct view about
the relative values of different functionings” (Sugden 1993, 1961).
Because it is “fuzzy” we need a lot of informational resources available
since no single perspective can adequately grasp all the various aspects

69
Apart from straightforward examples such as exclusively welfarist accounts of well-
being that see sick people as well if they seem happy with their condition, or the
compatibility of libertarian justice with famines, Sen shows that even more nuanced
less tendentious systems of closed evaluation, such as the bargaining framework of
Rawlsian social contract theory, are problematic (e.g. Sen 2002d).
 65

that seem involved. And we have to accept that this is not an

intermediate stage, on the way to finding the right theoretical
perspective under which it all snaps into sharp focus. The pluralism is
constitutive and not an embarrassment (Sen 1999a, 77).
This does not at all imply a rejection of theory, including highly
abstract ‘ideal’ theory. It has two important roles. Firstly, formal
deductive reasoning can be useful in extending and improving our
common-sense intuitions, and thus improves our understanding of
issues relevant to practical decision-making (Sen 2012a, 106). As Sen put
it in his Nobel lecture on social choice,

Informal insights, important as they are, cannot replace the formal

investigations that are needed to examine the congruity and cogency
of combinations of values and of apparently plausible demands. (Sen
1999b, 353)

Sen defends the highly abstract theorising of social choice (to which
he has made many contributions of his own) in part because of the
impossibility results it tends to generate. These show analytically that
certain common sense or commonly held moral intuitions cannot be
realised at the same time (such as Pareto optimality and minimal liberty
(Sen 1970a)).70 Sen has also shown analytically that rationality does not
require a complete ordering (Sen 1993c), and that the comparative
evaluation of states in terms of justice does not require any reference to
an ideally just state of affairs (Sen 2006b).
Secondly, theories provide us with interpretative concepts. They
identify certain features as important, justify that focus, and develop its
implications. Within ethics for example, libertarianism provides a
forceful argument for taking a certain class of individual property rights
seriously; utilitarianism for the significance of each individual’s
subjective mental states; and so on. Such theories enrich our conceptual
repertoire by telling us more about certain aspects of the world, why
they matter, and how they work. A pluralist account like Sen’s can make
use of this theoretical work (as a free rider) by using the accounts as

70
As Sen explains this ‘negative’ use of theory, “We often think, if only implicitly, of
the plausibility of principles in a number of specific cases . . . But once the principles
are formulated in unconstrained terms, covering inter alia a great many cases other
than those that motivated our interest in those principles, we can run into difficulties
that were not foreseen earlier, when we signed up, as it were, on a dotted line. We then
have to decide what has to give and why.” (Sen 2009a, 107)
66

observational vantage points to toggle between, as working objects that

can be more or less useful in improving our understanding of the issue
on which they are brought to bear.
The fact that various theories may plausibly be relevant does not
mean that all are equally relevant to the assessment of a particular case.
On the contrary, it raises the significance of the purposes and
perspective for which and from which a particular evaluative
assessment is done. The evaluator must take seriously – and take
responsibility for - the choice of techniques underlying an assessment
and not merely their application. It should not be an embarassment to
the social scientist to admit this and make the justification of her choice
of techniques explicit so that their suitability can be debated. In this
line, Sen argues against technocratic approaches to assigning weights to
different capabilities that avoid “messy” democratic procedures in
favour of some ready-made weights, for example as “revealed” by
market exchange value (Sen 1999a, 76–81). Such approaches constitute
undeclared value judgements by social scientists, and to the extent that
they fail to make their valuational perspectives explicit, impede rather
than support public scrutiny and deliberation. As I will now turn to
showing, for Sen the objectivity of social science depends on its
commitment to evaluation rather than valuation.

b. The Objectivity of Social Science

“Poverty, like beauty, lies in the eye of the beholder.” Does it really?
(Sen 1980, 366)

Sen has been called the “conscience of economics” by Robert Solow

and lauded by Hilary Putnam and Vivian Walsh for his revival of the
classical economists’ concern with ethics (Putnam 2002; Walsh 2008).
Sen certainly does believe in a two way relationship between ethics and
economics, since much applied ethical analysis depends on an
understanding of the social logistics that economics specialises in, while
economic policy-making is in turn dependent on the identification and
justification of values that is the specialism of ethics (Sen 2007). But his
view of the relationship is nuanced, and he vigorously maintains the
importance of maintaining a distinction between facts and values.71

71
A distinction is, of course, rather different than the metaphysical dichotomy that
entered economics with logical positivism (claiming its origins in Hume’s is-ought
distinction) which Putnam has argued to be untenable (Putnam 2002).
 67

Specifically, Sen has written in defence of the objectivity of the social

sciences in terms of the independence of the exercise of evaluation (in
which norms and values may be relevant) from that of valuation itself
(e.g. the claims that poverty is just inequality) (e.g. Sen 1980; Sen 1983a;
Sen 2004b). While evaluation concerns the assessment of the extent to
which valuable objects are achieved, valuation concerns the
determination of which objects are to be considered valuable (cf the
earlier discussion in chapter 1:II). Sen argues that what social scientists
like economists should do is evaluation not valuation, and that it is an
abrogation of professional responsibility to elide the two. While “facts
are endless, and to decide what to look at in empirical economics, we
need to have a sense of what is important and valuable” nevertheless,
“once our ethics lead us to the kind of questions that we ought to ask,
we must then seek, in the case of mainly empirical questions, as
factually sound answers as possible” (Sen 2003a, 327–8).
Sen’s criticism of the popular view that poverty is a value judgement
is illustrative here. This view was espoused for example by Mollie
Orshansky, who developed the official government poverty thresholds
still used in America to identify and count the poor.

For deciding who is poor, prayers are more relevant than calculation
because poverty, like beauty, lies in the eye of the beholder. Poverty
is a value judgment. (Orshansky 1969, 37)

Sen criticises this view as “an amalgam of taking a prescriptive view

of description (of poverty, in this case), and - additionally - a subjectivist
view of prescription (Sen 1980, 366).” That is, the definition of “poverty”
is a value judgement, and value judgements are matters of opinion. Sen
characterises such an exercise as “primarily a subjective one, unleashing
one's morals on the statistics of deprivation” (Sen 1979e, 285; Sen 1980,
366). What Sen is concerned with here is the direct perception of a case
under the terms prescribed by a particular valuational perspective
without going through the intermediate process of trying to grasp it in
its own terms, to describe it as it is. A value judgement is thereby
substituted for an evaluational judgement.
Instead Sen argues for seeing the assessment of poverty – and other
ambiguous and partially opaque concepts like well-being or inequality -
not as a value judgement or other subjective exercise but as “an exercise
of description assessing the predicament of people in terms of the
prevailing standards of necessities” (Sen 1979e, 287 emphases added).
68

The task is to make a second order professional judgement about the

extent to which various first order standards or values are in fact met.
Those standards are social ones and therefore reflect social value
judgements, but this doesn’t mean that describing how the state of
various members of society measures up on these standards is itself a
value judgement in the sense of signing up to those standards (Sen
1992, 107–9). Rather, it is an empirical, factual exercise. Social scientists,
even philosophical ones, are not supposed to decide, of themselves,
what is and is not valuable, nor even how much. That is not where their
expertise or authority lies. What they are supposed to be experts in is
the methodology of evaluation.
This section has discussed Sen’s explicit commitments to open and
objective evaluation and the role they play in his social scientific
methodology of evaluation. The concern is not with how the world
should be (as in mainstream justice theorising) but with understanding
how the world in fact is. Values, ideals, and ethical theories, play a
derivative role in this as standards relative to which how the world is
doing may be evaluated. Choices about which standards to employ and
how are matters of judgement by the observer, which must be explicitly
stated and justified rather than asserted as a matter of personal
opinion. In the following two sections I show how Sen’s concern with the
methodology of evaluation drives his contributions in two areas not
directly related to the capability approach: his critique of the
“transcendental institutionalist” approach to thinking about justice, and
his entitlement approach to food security.

III. SEN’S CRITIQUE OF TRANSCENDENTAL INSTITUTIONALISM

There is a strong case....for replacing what I have been calling

transcendental institutionalism – that underlies most of the
mainstream approaches to justice in contemporary political
philosophy, including John Rawls’s theory of justice as fairness – by
focusing questions of justice, first, on assessments of social
realizations, that is, on what actually happens (rather than merely on
the appraisal of institutions and arrangements); and second, on
comparative issues of enhancement of justice (rather than trying to
identify perfectly just arrangements). (Sen 2009a, 410)

Sen is extremely sceptical of what he calls the “transcendental

institutionalist” approach to justice which “looks for an ideal blueprint
 69

of social arrangements that cannot be transcended” (Sen 2008b).72 This

term brings together two distinct concerns that Sen has about the
dominant contemporary approach to theorising about justice. First its
‘transcendental’ nature which consists of a style of ‘ideal theorising’
concerned with identifying the completely ideal conditions of a perfect
end-state (which cannot be transcended). Second, its focus on
institutional arrangements, on identifying, albeit in different ways, the
perfectly just order of a society, without much regard for the overall
nature of that resulting society or the nature of actual, imperfect human
beings. Although these two features are distinct, Sen argues that they
are generally found together in the approach to political philosophy
exemplified by Thomas Hobbes, John Locke, Jean-Jacques Rousseau, and
Immanuel Kant. It is this same approach, he argues, that has been
predominant in twentieth century moral and political philosophy, most
prominently in the work of Rawls. It has come to the point that ideal
theorising about institutions is taken for granted as the necessary
starting point for discussions of justice (as in the operationalist
critiques of Pogge and Nussbaum).
Sen argues instead for a ‘comparative’ approach to justice which
relates to his work on social choice theory and his understanding of its
traditional concerns. He relates this to a tradition of realisation-focused
comparative thinking about justice he finds in the work of the Marquis
de Condorcet, Adam Smith, Jeremy Bentham, Thomas Jefferson, Mary
Wollstonecraft, John Stuart Mill, and Karl Marx (though they had quite
different interpretations of the basis of justice comparisons) (Sen 2009a,
xvi). The idea is to focus pragmatically on identifying the agreements
that exist across ethical and political perspectives, particularly about
what all agree are significant injustices that should be removed. Sen
argues that partial orderings based on dominance or the intersection of
different individual orderings or valuational principles can, even without
debate, rule out many ‘extreme’ choices, such as the institution of
slavery.73

72
Sen has elaborated on this critique in a series of recent publications (see for example
Sen 2006b; Sen 2008a; Sen 2009a).
73
There are many ways of deriving such partial orderings, for example interpersonal
preference rankings or different ethical theories – we don’t have to decide on a
particular ethical theory or social choice theory account before the analysis if we find a
considerable intersection of their rankings with respect to a particular issue. Sen also
draws on Rawls’ concept of public reason as a way of extending such partial orderings
through the consideration of generally acceptable (public) reasons. I elaborate on Sen’s
social choice approach to social justice in chapter 6.
70

The emphasis is on making things better (amelioration) rather than

perfect. In terms of institutions, society can come to partial agreements
based on the overlapping of conclusions supported by different
reasoning. Certain institutions can be generally agreed to be gravely
unjust, such as slavery or the legal status of women in marriage in
nineteenth century England, without agreeing on what a perfectly just
society would be like. The banning of slavery and the expansion of
women’s legal rights clearly addressed urgent injustices and made
society more just (a comparative improvement). But they were not
collectively motivated by, and arguably got no closer to, a shared vision
of an ideally just society (Sen 2009a, ix–x). Such comparative
improvements are of course further revisable in the light of changed
empirical or ethical understandings and political feasibilities.74
For the purposes of this section I will not attempt to adjudicate the
extensive and wide-ranging contemporary debate about the place of
‘ideal theory’ in political philosophy (Robeyns 2008; Valentini 2012),
which Sen’s high profile intervention has particularly prompted.75 This
should not be taken to imply neutrality, since I do find Sen’s position
very persuasive. I will make one small intervention, which I haven’t seen
elsewhere, to show that the implications of Lipsey and Lancaster’s
General Theory of Second Best provide further support for Sen’s case,
and also emphasise the role of judgement.
Sen’s criticism of transcendental institutionalism is methodological.76
Here I consider three distinct problems he raises. First, there is the idea
that thinking about justice requires identifying a complete ordering of
every state before any decision about what is the best or right thing to
do can be justified. (This was Pogge’s claim in setting up his critique of
Sen, discussed in the previous chapter.) Sen has spent decades showing

74
There is much in common here with the cautious boldness recommended by J. S.
Mill, for example in his Considerations on Representative Government (Mill 1861) in
which he argues both against committing to radical ‘progress’ and being satisfied with
the injustices of the status quo. See for example Sen’s qualified defence of the World
Bank and IMF in his review of William Easterly’s White man’s burden (Sen 2006c) or the
article on development that he co-wrote with the president of the World Bank (Sen and
Wolfensohn 1999).
75
Sen is not the only such critic, of course (see, among others, Wolff 1998; Mills 2005;
Farrelly 2007). It is also worth noting that the points he raises are quite specific to the
‘transcendental institutionalist’ version of ideal theorising and thus somewhat
orthogonal to that ongoing debate.
76
As John Davis notes, “Sen does not begin [The Idea of Justice] with the question ‘what
theory of justice?’ – a substantive sort of concern – but rather begins with the
question, ‘what kind of a theory?’ – a methodological sort of concern.” (Davis 2012,
169)
 71

that a complete ordering is not required for rational choice between

alternatives and may in fact be counter-productive. To give a common
example of Sen’s, Buridan’s ass was supposedly so focussed on deciding
which of two haystacks was the best, that it neglected the possibility of
making a good decision and starved itself by failing to choose at all (Sen
2000a, 487). With respect to more immediate issues, such as gender
equality, we do not need to know what a society with true equality
between men and women would look like in order to criticise many
existing institutions and practices (such as in the division of labour in
care and household work) as grossly and unnecessarily unfair (Sen
2012b, 175).
In fact, the pursuit of a complete ranking can distract us from the
considerable scope that may already exist for making some reasoned
judgements with the information and resources we do have, about
identifying and addressing pressing problems. Sen identifies two ways
of doing so: the principle of maximisation and the partial-intersection
approach. Maximisation, as distinct from optimisation (the pursuit of
the first-best option), merely says that one should always choose an
option that is not known to be worse than any other. The partial
intersection approach concerns finding points or ranges of agreement
between different rankings of states of affairs, such as between the
different rankings that different theories of justice produce. This can
allow us to judge various states of affairs – such as slavery - as
definitively unjust without first deciding which theory of justice is
correct.
A second problem with the transcendental institutionalist approach
is that it can distract us into a very abstract theoretical world of puzzle-
solving at the expense of paying due attention to the important features
of real world issues. The goal of generating a complete ordering may
well require a narrow theoretical framework that achieves its goal by
idealising away (abstracting from) significant aspects of a situation, such
as inequality (utilitarianism) or disability (Rawls’ social contract). Sen
argues that while such theoretical constructs may achieve their goals of
achieving internal consistency and providing precise answers about
what justice requires, this is often achieved by deliberately
impoverishing the scope of the analysis (see e.g. Sen 1999a, chap. 3).
Thirdly, the specification of first-best institutional arrangements
lacks real world relevance since it simply presumes the ideal conditions
that it requires. It does not provide a clear goal or even a road map
72

towards its achievement in a second-best world. Sen argues that the

transcendental perspective is neither necessary nor sufficient to help us
answer questions about comparative justice – what should we do in this
world to make it less unjust – that he takes to be the important ones
(Sen 2006b). As Sen puts it,

A transcendental approach cannot, on its own, address questions

about advancing justice and compare alternative proposals for
having a more just society, short of proposing a radical jump to a
perfectly just world. Indeed, the answers that a transcendental
approach to justice gives—or can give—are quite distinct and distant
from the type of concerns that engage people in discussions on
justice and injustice in the world, for example, iniquities of hunger,
illiteracy, torture, arbitrary incarceration, or medical exclusion as
particular social features that need remedying. The focus of these
engagements tends to be on the ways and means of advancing
justice—or reducing injustice—in the world by remedying these
inequities, rather than on looking only for the simultaneous
fulfilment of the entire cluster of perfectly just societal
arrangements demanded by a particular transcendental theory. (Sen
2006b, 218)

Some have argued that Sen’s argument for the redundancy of

theories of perfect institutional arrangements goes too far. For example,
Ingrid Robeyns argues that such theorising is necessary, though not
sufficient, for identifying and ameliorating injustice, exactly because it
provides a picture of what it is one is seeking to achieve (Robeyns 2008;
Robeyns 2012). Indeed, it seems obvious that having the blueprint for
the design of the perfectly just society in one’s hand must be not only
helpful but essential for building such a society, and also for identifying
how our present society’s arrangements differ from what perfect justice
requires.77 Yet, if one switches metaphors, it is equally obvious that
knowing what one’s destination looks like (in terms of its institutional
arrangements, as in Thomas More’s island of Utopia) does not in itself
provide any information about how to get there.78 I think this shows the
limitations of arguing through metaphor.

77
This seems to have been Rawls’ own understanding of the role of the kind of ideal
theory he was advancing: “The reason for beginning with ideal theory is that it
provides, I believe, the only basis for the systematic grasp of […] more pressing
problems” (Rawls 1999, 8).
78
The Latin components of the word ‘Utopia’ suggest that it can be read both as ‘good
place’ and ‘no place’. Certainly More’s actions as Lord Chancellor of England did not
 73

Here an analytical result from economics about the place of ideal

theory seems relevant. In their paper on The General Theory of Second
Best, Lipsey and Lancaster demonstrated that the standard approach of
attempting piecemeal welfare economics improvements was wrong. This
approach consisted of trying to come closer to the optimum conditions
required for the attainment of a Paretian optimum in any particular area
of policy. For example, whenever one comes across a trade-barrier or
monopoly that one can remove, one should do so. It was based on the
assumption that the closer arrangements in the real world resembled
the conditions specified in the theory, the closer the performance of the
economy would get to optimal efficiency.
It turns out, however, that once one departs in any irremediable
respect from the conditions assumed by the ideal theory, then one is in
a non-ideal situation.79 In this situation one must seek a second-best
(rather than the first-best) set of arrangements. Unfortunately, in
identifying what those arrangements would be the first-best rules are
irrelevant. First, they can provide no positive guidance for solving the
complicated multi-dimensional problem of identifying a second-best set-
up (a rather more difficult problem than that for which they were
designed). Second, the only thing one can know for sure is that acting as
if they did provide guidance – trying wherever possible to imitate the
conditions required for the first-best optimum - would generally be
unsuccessful.80 Lipsey concludes that,

Practical policy advice requires more parochial objective functions

than community welfare; must rely on formal and appreciative
[verbal] theory, empirical evidence, and large doses of judgment; and
should concentrate on making piecemeal improvements in context-
specific situations. (Lipsey 2007, 349 emphases added)

This result would seem applicable to the kind of transcendental

institutionalist theory Sen criticises. The blueprints for an ideal society’s

suggest that he saw the arrangements he identified as Utopian, such as freedom of

religion, as providing direct instructions.
79
“The general theorem for the second best optimums states that if there is introduced
into a general equilibrium system a constraint which prevents the attainment of one of
the Paretian conditions, the other Paretian conditions, although still attainable, are, in
general, no longer desirable. In other words, given that one of the Paretian optimum
conditions cannot be fulfilled, then an optimum situation can be achieved only by
departing from all the other Paretian conditions.” (Lipsey and Lancaster 1956, 11)
80
For example “the adoption of a free trade policy by one country, in a multi-country
tariff ridden world, may actually lower the real income of that country and of the
world” (Lipsey and Lancaster 1956, 14).
74

institutions are not going to be relevant if one doesn’t have all the
materials one needs to build it. For example, the entitlement theory
expounded by Robert Nozick in Anarchy, State, and Utopia (1974) is
often interpreted as recommending realising libertarian arrangements
for justice in acquisition and transfer. However, as Nozick notes, it is an
historical account. Thus, it requires the specification and enactment of a
principle of rectification to fix all historical violations of justice in
acquisition and justice in transfer (Nozick 1974, 150–3).81 Unless this
requirement for rectification is met (and it seems utterly implausible
that it can be), there is no basis for assuming that institutionalising the
requirements of justice in transfer in the present (for example by
protecting existing property holdings from redistributive taxation) will
bring us any closer to true libertarian justice. A similar point might be
made of Rawls’ assumption of standard endowments. As long as there is
substantial heterogeneity in people’s ability to convert resources into
valuable functionings, giving people equal resources cannot be
presumed to enhance fairness in the sense of giving everyone an equal
opportunity to live a life according to their conception of the good. Two
people with the same conception of the good life might well have quite
different real opportunities to advance it (Sen 1990a).
This is not to say that the ethical reasoning behind the selection of
those institutional arrangements will not remain pertinent (cf Swift
2008). For example, an ideally just society would presumably be one in
which racist and sexist discrimination was not permitted i.e. it would be
characterised by an institutional rule of ‘no discrimination on the basis
of race or gender’. Yet there such discrimination appears endemic in all
existing societies. Therefore, the same ideal of enhancing substantive
equality may justify a deviation from ideal institutional arrangements
(the ‘no discrimination’ rule) such as some form of ‘affirmative action’
policy or programme.82 Whether or not an affirmative action policy is
justified, and if so of what kind, is, however, highly contingent on the
character of that situation and is a decision that requires judgement
81
Nozick’s famous Wilt Chamberlin example obscures this condition, since it
introduces a year zero.
82
This is a contemporary debate, for example in America where affirmative action
programmes are often challenged on the grounds that they violate laws against racial
discrimination. Chief Justice Roberts, in ruling unconstitutional a school board’s use of
racial classification in assigning students to schools, concluded, “The way to stop
discrimination on the basis of race is to stop discriminating on the basis of race”
(Roberts 2007, 41). This case is also discussed by Nussbaum in her analysis of
constitutional law and the capability approach, and her critique of “lofty formalism”
(Nussbaum 2009a)
 75

rather than the direct application of a rule identified by a priori

reasoning.
This section has explored Sen’s methodological argument against the
relevance of transcendental institutionalist theories of justice for
thinking about how we might make this world more just. The
interpretation I have given of this critique is that one cannot escape the
need for contextual and multi-dimensional evaluative exercises and that
these require judgement rather than the direct application of the
‘correct’ theory. The next section considers a somewhat more practical
example of Sen’s methodology of evaluation, his entitlement approach
to food security.

IV THE ENTITLEMENT APPROACH

Sen’s entitlement approach provides a contrastive example to the very
abstract critique of transcendental institutionalism just considered. It
prefigures the capability approach, and can be seen as a simplified
version of it in which certain features of Sen’s evaluational methodology
may be highlighted. It concerns a single capability (having enough to eat)
and limits itself to considering legal entitlements for this. Nonetheless
the combination of Sen’s disaggregated evaluative methodology with
loose organising concepts developed to suit it has dramatically changed
how we understand and seek to prevent famines.
Sen has severely criticised what he terms the Food Availability
Decline (FAD) approach as mistaken and dangerous (Sen 1981; 1982). It
is mistaken because many historical famines have in fact been caused or
exacerbated by factors other than declines in the quantity of locally
available food (such as a collapse in trade entitlement in an exchange
economy). It is dangerous because adherence to the FAD understanding
of famines has led to egregious mistakes in the identification and policy
responses to famines (and pre-famine conditions), particularly with
regard to the impact of famines on different groups within the same
population.
For example, Sen’s analysis suggests that the Bengali famine of 1943,
which killed perhaps 3 million people, was principally caused by a
collapse in rural food purchasing power (particularly to agricultural
labourers) due to wartime inflation and British policies, including a
wartime information blackout and general disinterestedness in
circumstances outside cities (Sen 1981, 441–447; Sen 1982, 52–85). It
was thus a ‘boom famine’, caused not principally by supply side shocks
76

(FAD) but by demand side forces: the uneven expansionary economic

forces which left out certain occupational groups and thus induced a
collapse in their ability to buy sufficient food.
However Sen does not argue against the FAD hypothesis - that
famines can be caused by FAD - only against the claim that FAD is a
necessary, rather than merely a sufficient cause of famine. The FAD
causal hypothesis is instead subsumed within Sen’s own much broader
entitlement approach as something that might affect individuals’
“entitlement” to food through various mechanisms. This entitlement
approach, Sen suggests, is the best way to understand all famines.
People have endowment sets consisting of their legally owned resources,
and final entitlement sets consisting of the resources they can legally
obtain using their endowment set, while the entitlement mapping is the
relationship between the two sets (and includes for example,
production, exchange and transfer components) (Osmani 1995, 254–5).
This provides us with the apparently absurdly simple diagram:

Endowment  Entitlement
Entitlement mapping

In his excellent analysis of Sen’s work on the entitlement approach

(Osmani 1995), Siddiq Osmani explains the significance of what might
look like a definitional sleight of hand (or to Pogge merely “a new
language”). Sen’s entitlement approach “purports to advance no causal
hypothesis, only an organizing framework within which various causal
influences can be systematically explored” (Osmani 1995, 264 emphases
added). By redefining famines as mass entitlement failures (for which
FAD may be a causal factor), Sen places individuals at the centre of the
analysis (Osmani 1995, 269). Starvation, as Sen forcefully points out, “is
the characteristic of some people not having enough food to eat. It is
not the characteristic of their being not enough food to eat...Starvation
statements are about the relationship of persons to the commodity (Sen
1982, 1 original emphases).” One can see the continuity with the
capability approach here, in Sen’s focus on the relationship between
individuals, commodities, and valuable achievements.
The entitlement approach proceeds by asking, ‘do individuals here
have sufficient entitlement to food?’ (and not ‘how much food is there in
the area?’). If an entitlement failure is diagnosed among some people,
then an explanation of its determinants is sought across the various
 77

interconnected causal mechanisms affecting their endowment and

entitlement mapping, which might for example include rural food price
inflation (which particularly affects groups such as landless labourers,
as in Bengal in 1943) or a decline in food production (such as caused by
a drought) (Sen 1982). The diagnosis of an entitlement failure thus
motivates, and its particular nature directs, a search for causes affecting
the endowment sets and entitlement mappings of different (particularly
socio-economically distinguished) individuals (Osmani 1995, 269). The
disaggregated nature of this analysis and the flexible conceptual
framework of the entitlement approach allows an open-minded search
into the causal mechanisms relevant to any particular case which can
often turn up unexpected relationships (as Sen illustrated with his
analysis of four very different historical famines (Sen 1982)), while at
the same time discriminating the asymmetric impact of such forces on
differently situated individuals (Osmani 1995, 269).83
The entitlement approach focuses on the methodology of evaluation.
Just as with the capability approach, it begins from a critique of the
standard way of understanding famines, as a sudden shortage of food.
That approach neglected the crucial aspect of whether actual individuals
were able to get sufficient food. Thus Sen introduced this relationship
as the core of his framework for understanding famines. Yet the simple
organising framework he proposed is an open one which permits, and
requires, the consideration of many disparate causal factors. In
particular, although the framework itself is quite simple, the scope of
the empirical analysis it requires is much broader than its mono-causal
predecessor, in which one simply counted up the total amount of food
and divided by the number of people.84 The entitlement approach

83
Note that just as the capability approach is often mistaken for a theory, or confused
with a particular capability theory, Sen’s entitlement approach is still frequently
confused with being an entitlement hypothesis (e.g. as claiming that famines are always
caused by falls in relative purchasing power), and thus being a direct competitor with
FAD hypotheses. Particular entitlement hypotheses can be opposed to FAD ones and
tested against the empirical evidence without prejudice to the assessment of Sen’s
entitlement approach (see for an example Lin and Yang 2000). Likewise, the
comparative assessment of capability and non-capability theories need not have any
implications for the capability approach itself. Regarding policy, Sen himself has
endorsed greater use of transfer payments (e.g. employment programmes) to shore up
entitlement sets and prevent famine (which parallels his endorsement of a shortlist of
basic capabilities for general development purposes) but nevertheless maintains that
the wider evaluative framework always remains relevant to the analysis of a particular
situation.
84
Though Sen is careful to note the limitations of his analytical framework: real-world
entitlements may be ‘fuzzier’; extra-legal command of food (i.e. outside the entitlement
framework) may be significant; people may consume less than their legal entitlement;
78

requires a much more disaggregated empirical analysis that goes

beyond counting. For if the central question is the relation between
individuals and food (rather than populations and food), then the
different abilities of differently situated individuals to get access to food
must be discriminated and given specific attention.
How we see the world matters. “A misconceived theory can kill, and
the Malthusian perspective of food-to-population ratio has much blood
on its hands.” (Sen 1999a, 209) The persistent misunderstanding of
famine long into the modern world led to millions of easily avoidable
deaths. While FAD treats famines as natural phenomena, Sen’s
entitlement approach allows them to be seen in broader terms, as a
failure of a society’s political-economy arrangements to provide security
of entitlement to all. That characterisation supports political action,
engaging people as agents in demanding government interventions
rather than as passive victims.85 The entitlement approach, like the
previous example and the preceding analysis of Sen’s methodology,
demonstrates the centrality of evaluation to his work. In the following
sections I will turn to analysing Sen’s concern with gaining an adequate
understanding of the world in philosophical terms.

V. FLEISCHACKER: REFLECTIVE AND DETERMINATE JUDGEMENT

So far I have explored Sen’s methodology of evaluation without
attempting a philosophical analysis. It is time to provide one. Sen’s
foundational commitment to evaluation can best be understood, I will
argue, in terms of judgement. I am using the term judgement in quite a
specific sense, following Samuel Fleischacker’s (1999) conceptual
analysis of ethical judgement. I will first discuss Fleischacker’s account
and then in the following section the implications of this way of seeing
Sen’s account of the capability approach.
Fleischacker argues for distinguishing judgement from other
intellectual skills such as “perceptiveness” or “analytical rigour”. It is
particularly significant in making decisions under uncertainty, where the
uncertainty concerns “what canons of success it is appropriate to use”

the high mortality associated with famines is principally caused by epidemics, which
must be analysed separately (for example as due to refugee movements and
breakdowns in sanitation as well as starvation-weakened immune systems) (Sen 1981,
437–439). For a critical analysis of the contribution of Sen’s entitlement approach see
Des Gasper (Gasper 1993).
85
Sen argues the relative ease with which an adequately informed and motivated
government can prevent famines (which rarely affect more than 5% of the population),
explains why they are absent in democracies (Sen 1999c, 7–9; Sen 1999a, 160–188).
 79

(Fleischacker 1999, 16 original emphases). That is, where we have to

work out, as part of the exercise of making sense of something, what the
criteria for such an understanding should be. This fits well, I believe,
with what I have been claiming about Sen’s concerns in evaluation.
Fleischacker derives his account of how judgement works principally
from his reading of Kant’s analysis of aesthetic judgement in his
Critique of Judgement, complemented with his reading of Adam Smith.
He characterises judgement as a dynamic two phase process that
toggles between two interdependent but incommensurable phases in
search of the best way to understand an issue before you. Reflective
judgement is the search for the right general concept for grasping a case
(which in ethics, for example, might include various conceptions of
freedom). Determinate judgement is the application of a general concept
to a particular case, to see it as one kind of thing rather than another (as
liberty or autonomy, for example). The relationship between the two is
described by Kant in terms of “the free play of the imagination and the
understanding”. 86 More specifically,

Judgment in general is the capacity to think the particular as

contained under the general. If the general (the rule, the principle,
the law) is given, then the judgment that subsumes the particular is
determining....If, however, only the particular is given, for which the
general must be found, then the judgement is merely reflective.

Determining judgements are conclusions about how something

should be understood, as one thing rather than another. That is, they
consist of the application of one particular concept, rule, or principle to
the object under consideration. Determining judgement is particularly
dominant (that is, at the expense of reflective judgement) in disciplines
in which fundamental concepts are not usually reflected upon. For
example, much of the natural sciences, most of the time, are
characterised to a high degree by determining judgement in which the
central problem is the application of a standard theoretical account to
new particular cases (data) (Fleischacker 1999, 28). Nevertheless even in
such disciplines the dominance of determinate judgement is not
categorical but rather a matter of degree. There is always at least the
possibility of reflective judgement remaining in the background, and the
routine application of the standard theoretical framework can always be

86
This quote, and the following, are from Kant’s Critique of Judgement, translated and
quoted by Fleischacker (pp 23 and 12, respectively)
80

challenged for not being properly ‘responsible’ or adequate to the

nature of the particular case (as Sen achieved with his entitlement
approach to food security, for example).
In contrast, reflective judgement is concerned with the imaginative
exercise of trying to consider what kind of concept would be adequate
for understanding a particular case that eludes our immediate
understanding. Reflective judgement thus requires exploring whether
and how various concepts, principles and rules might adequately
organise our particular intuitions about the case at hand. There is a high
degree of playfulness in this exercise, as different perspectives are tried
out, discussed, reworked, and discarded in an indefinitely extended
process.
Nevertheless, reflective judgement can be distinguished both from
merely “musing” over particulars or generalities without coming to
conclusions (involving feelings but not cognitive content) and abstract
theorising (in which general conclusions are deduced from other
generalities) (Fleischacker 1999, 8–9). It is a mode of thought
particularly dominant in disciplines where “conclusions depend in
significant part on grasping the features of complex and unreproducible
particular cases”, such as aesthetics, literature, and the social sciences
(Fleischacker 1999, 9).
Reflective and determinate judgement are different sides of the
same coin: where the former is engaged with considering the
appropriate concepts, categories, and rules of interpretation, the latter
is concerned with their application (Fleischacker 1999, 13–14). In the
absence of reflective judgement, all one would be able to do is apply the
same concept to everything. In the absence of determinate judgement
one lost in a sea of concepts unable to come to any conclusions about
what one is looking at.
Two further features of Fleischacker’s account are also relevant. The
conclusion of an exercise of judgement, of deciding how a particular
case should be understood, remains defeasible rather than definitive. It
can be re-opened and its basis reconsidered, for example in the light of
new information. Another feature of judgement is its inter-subjectivity.
Unlike the expression of an idiosyncratic opinion, judgement is an
exercise of impartial reasoning – for example, about which aspects of a
case to treat as salient and how to weight them in order to best
understand it - and that reasoning should in principle be justifiable to
others. Other interested parties are able to, and implicitly invited to,
 81

consider your reasoning for themselves. While there is no requirement

that others agree with your conclusion, even where they have the same
information as you, there is an implicit requirement that they at least
consider the way you came to your conclusions to be reasonable. The
legal judge whose judgements are founded on racial bias is quickly
identified and condemned. Although judgement is carried out as a
private mental exercise, its goal of coming to the best conclusion carries
the implication that others would also, perhaps after suitable
discussion, either see your point or persuade you better.
It is worth noting the parallels and contrasts of this account of
judgement with the account of “judicious spectatorship” developed by
Martha Nussbaum (Nussbaum 1995; Nussbaum 2009a). Nussbaum’s
account elaborates an aspect of practical reason she calls “analogical
reasoning” as a model for legal case analysis. The judicious spectator is
recommended to attend to individuals’ particular circumstances as one
would read a novel. This has many superficial similarities with
Fleischacker’s account (even including references to Adam Smith’s
impartial spectator). Nevertheless it differs significantly both in form
and orientation.
First Nussbaum focuses on the direct “perception” aspect of
practical reason at the expense of the development and choice over
abstract concepts and rules (Fleischacker 1999, 33).87 She wants us to
train our mental vision in certain ways (literature is particularly
important) so that we attend empathetically to the story of people’s
lives that a legal case represents. This emphasis on direct perception
comes at the expense of the fuller exercise of reflective judgement, of
considering the relevance of various alternative and more abstract
conceptual resources from which objections and corrections to a
misguided immediate perception of a case can be raised.
For example, in her use of the narrative method to motivate her
capability theory of justice Nussbaum frequently recounts the heart-
wrenching story of Vasanti, an illiterate impoverished Indian woman
rescued by a women’s charity (Nussbaum 2000, 15–24). But she has been
criticised for over-interpreting the voice of her interviewees in order to
make them fit into her argument. Susan Okin, for example, has argued
that, rather than being in a dialogue with these women, “Almost

87
Thus, “The CA, I hold, needs judges who reason in a certain way, with a certain
quality of mind: Aristotle, describing the virtues of such a judge, called it ‘perception,’
contrasting it with a mere deference to rules”. (Nussbaum 2009a, 341)
82

everything Nussbaum says about the two women, their lives, and even
their thoughts, perceptions, and emotions is filtered through her, and
much of is prefaced by phrases like “it seems” or “suppose” (Okin 2003,
295).”
Second, and relatedly, Nussbaum’s use of this technique is rather
one sided. In any particular case she only directs the judicious spectator
to consider the real situation of whichever party appears to be the
‘underdog’.88 Thus, wherever Nussbaum employs her technique of
judicious spectatorship on legal cases, she always tells only one
(nuanced and persuasive) story of the person she has identified as the
victim of injustice. One wonders whether this is because the power of
narrative would allow the ‘wrong side’ to make persuasive arguments
too. In contrast, in his examples Fleischacker presents the strongest
arguments of both sides, to show how tricky it is to say which legal
concept a case should best be understood under, and thus determine
the judge’s decision (e.g. Fleischacker 1999, 8–13).
The point of judgement, as Fleischacker presents it, is to try to make
the best decision about how to treat a case by considering which among
various competing framings of it best make sense of it. This is a difficult
and challenging task. Nussbaum, however, seems to think that deciding
who is right is actually quite straightforward. The problem is in
persuading the people who do the judging of the right ethical account.
But then Nussbaum’s account seems limited to the promotion of her
favoured ethical theory. She is interested in criticising certain kinds of
prejudices (against underdogs), but her method does not provide critical
distance from other kinds of prejudices one might have (including
presumptions in favour of the underdog).89 She does not endorse the
kind of experimental approach at the heart of Fleischacker’s account, of
trying to put one’s presumptions aside and look at a case in various
other ways. Thus, while Nussbaum’s ‘perception’ model may be a
powerful tool for advocacy, it is simply inappropriate as a model of
impartial judgement.

88
A point noted by others, such as Mary Sigler, who notes in her review of Poetic Justice
that, “In Nussbaum’s examples …. it almost seems that in order to decide whose story
is the relevant one, we must first determine who the sympathetic protagonist is.”
(Sigler 1997, 623)
89
As Nussbaum describes her approach in a recent interview, “I've always focused on
the underdog, and I don't like any form of beating up on the less powerful person....I
think the unity among the different issues I tackle is that I'm concerned with some
relatively powerless group that's getting stigmatized or beaten up on by other people.”
(Nussbaum 2009b)
 83

VI. IMPLICATIONS: THE ROLE OF THE MORAL OBSERVER IN THE

CAPABILITY APPROACH

The [capability] approach (sometimes Sen calls it a” perspective”)

does not pretend to yield a “decision method” that could be
programmed on a computer. What it does do is invite us to think
about what functionings form part of our and other cultures’
notions of a good life and to investigate just how much freedom to
achieve various of those functionings various groups of people in
different situations actually have. (Putnam 2002, 60)

By the ‘sleight of hand’ of defining advantage in terms of capability,

Sen provides a general organising framework within which the relevance
of various ethical and practical concerns can be systematically explored.
That exploration is a sophisticated exercise in evaluation that I think
falls within Fleischacker’s model of judgement. It certainly does not fall
within the model proposed by some critics of Sen (such as Pogge and
Nussbaum, discussed in the previous chapter). In that ‘operationalist’
exercise there is a search for the best concept of justice, but this is
understood as an exercise of analytical rigour, of theorising (“drawing
general conclusions from other generalities” as Fleischacker puts it
(Fleischacker 1999, 8)). There is also determinate judgement, since once
the right theory of justice has been identified its concepts, rules, and
principles are to be applied to all the cases one comes across. There is
little here of the sense of uncertainty that characterises the full exercise
of judgement according to Fleischacker, and that is so apparent in Sen’s
embrace of ambiguity and pluralism without “embarrassment” (Sen
1999a, 77). As Sen’s analysis of the informational bases of moral
judgements showed, the a priori exclusions of that approach mean that
one will always get the same kind of answer to every case (so one will
achieve consistency), but whether those answers are reasonable is
somewhat doubtful. To paraphrase Sen, the result is of unleashing one’s
morals on the world, rather than facing up to the challenge of properly
understanding the nature of the case in question.
Sen’s approach to evaluation takes that challenge seriously. That is
why he retains ambiguity both within the capability approach
(concerning which capabilities are valuable, and the relation between
functionings and capability) and without (such as the relevance of
procedural aspects like fairness and liberty). Making definitive decisions
84

about such matters would limit the conceptual resources available to the
evaluation of complex and opaque ideas like poverty and well-being.
Avoiding such difficult decisions is seen as evasive by some of Sen’s
critics, but it actually seems a requirement for the kind of critical
evaluation with which Sen considers himself concerned. (Indeed, from
Sen’s perspective it is those demanding complete theories who are
evading the requirements of the discipline of evaluation.)
Of course, any application of the capability approach, whether for
inter-personal comparison of advantage in some dimensions or
evaluating the nature of poverty in a rich country, will seek a
determinate judgement, and will therefore employ specific concepts and
related criteria. However, the flexibility of the underlying framework
means that this selection is the responsibility of the researcher rather
than the theory itself. As Sen puts it,

[A]ny description involves discrimination and selection, and the real

question is the relevance of the selection process to the objectives of
description. (Sen 1980, 361)

Thus, those using Sen’s capability approach are supposed to engage

in judgement in the full sense (both reflective and determinate aspects).
Yet a substantial criticism can be made of Sen’s emphasis on
evaluation, and thus on the judgement of researchers and policymakers
who take up his approach. By excluding the rule of theory, this central
role for judgement seems to introduce the rule of judges. This, I think,
underlies Robert Sugden’s concerns about the moral observer who
seems omnipresent in Sen’s work but whose role and accountability isn’t
properly explained.

Sen addresses his fundamental question [‘how a person’s interests

may be judged and his or her personal “state” assessed’] from many
angles, and takes account of many different aspects of well-being.
However, one feature that remains constant is the viewpoint from
which the question is posed – the viewpoint of a moral observer,
making ‘judgements’, ‘assessments’ or ‘evaluations’ of a person’s
state from outside. (Sugden 2006, 36)

The problem of depending on judges for capability evaluation is that the

methodological decisions they make about how to see cases can be
extremely important (for example with respect to shaping public
policies, and which people they are directed towards), yet, without an
 85

independent systematised methodology, it doesn’t seem possible to

hold them accountable. One must trust that they are experts and that
they are honest. Adherence to the kind of methodological guidelines
proposed by Robeyns (Robeyns 2005) for formulating and justifying
research selections, may play a significant role in warranting this trust. 90
Nevertheless, it remains the case that the quality of such justifications
can only be checked by other experts who thoroughly understand the
subject area (and not only the capability approach).
Even when the methodology is explicit, the ordinary reader can have
little understanding of the significance of particular decisions and their
sensibleness. For example, in 2007 UNICEF published a report card on
child well-being in rich countries (UNICEF 2007). Its attempt at
comprehensive multi-dimensional evaluation (incorporating 40 different
indicators relevant to children’s well-being and rights) is very much in
line with the spirit and methodology of the capability approach
However, the report was not only controversial in the positive sense that
it made uncomfortable reading for the citizens and governments of the
countries placed lower down the list. Its methodology immediately
became the object of controversy because of criticism by those who
didn’t like its results (particularly in the Anglo-Saxon countries that
came last). Why were certain things (like children not having to share a
bedroom) included? Why was relative income poverty weighted so
highly? And so on. The general challenge was that many of the criteria
on which the assessment was based were really matters of opinion
rather than facts about well-being.
The problem is that the more sophisticated the evaluation, the more
judgement is required and acknowledged in the methodological choices
made, and the harder it then is for the general public to understand and
scrutinise the underlying reasoning. There is thus an opacity in the
evaluation of capability, despite the intuitive accessibility of the basic
concept. Indeed, the controversy over the UNICEF report stands in
contrast to the public success of ‘vulgar’ figures like GDP per capita
rankings, which are perceived as being more objective because they are
not subject to the whims of researchers to define in a way that suits
their political agenda.
This may be the reason that Sen has emphasised, but not fully
explained, various features of the capability approach that increase the
accountability and diminish the authority of these moral observers. He

90
Robeyns’ guidelines were discussed in chapter 1: IV.
86

argues, for example, that the methodology behind ‘social evaluations’

should be open to public scrutiny (Sen 1999a, 78–81). He argues for
using existing social standards as the criteria for evaluation (though he
allows evaluators a choice over which social standards to use) (Sen
1979e, 285–6). And he argues for the foundational importance of what
the people concerned have reason to value (Sen 2008a, 231–2). In the
second half of this thesis these ‘democratic’ aspects of the capability
approach will be considered as I raise and address three distinct
challenges for the capability approach.

VII. CONCLUSION
The distinction between Sen’s approach and those he criticises is not
only substantive (i.e. that their concepts are inadequate), but also
methodological. In particular, to Sen, many seem over-determined with
regard to their subject matter. They ignore relevant information that
doesn’t have a place in their theory (as with utilitarians’ exclusive focus
on utility, or development economists’ economic metrics). They are too
quick to see a case one way and fail to even consider alternative
perspectives. They fail to acknowledge that the person doing the
evaluation is responsible for the choice of perspectives and not merely
for successfully applying the right theory.
The openness and yet objectivity of Sen’s capability approach – its
ability to consider the heterogeneity of individuals’ capability sets and
to be pluralist even in its principles of assessment – is essential to its
attractiveness as a flexible framework for evaluating advantage. But that
very openness is also seen by many as a problem that needs to be
overcome. In this chapter I have tried to show that Sen’s particular
methodology of evaluation is a constitutive feature of his approach and
has under-appreciated credibility and strengths.
Up to this point I have primarily been concerned with explicating
and justifying a distinct interpretation of Sen’s capability approach as a
“framework of thought”, and its place in the contemporary moral and
political philosophy literature. In the second half of this thesis I will be
concerned with what the capability approach can do. Each of the
following chapters considers the application of the capability approach
to a specific issue at a different level: practical reason (4), development
(5), and social justice (6). Each chapter is organised around a challenge
and contribution framework. I start by considering a significant
challenge that can be made against the capability approach (respectively,
 87

‘adaptive preferences’, paternalist development, and a political

philosophy gap). I then show that the capability approach (sometimes
suitably extended but in line with Sen’s concerns and methods) can meet
each of these challenges, and that this exercise also identifies and
clarifies several of its specific contributions and limitations.
88

Chapter 4. Sen’s Adaptive Preferences and Adam Smith’s

Impartial Spectator

Who are you, Justice Commissar, to say the Bengali beggar's

capability is less rich than the Princeton professor's? (Roemer
1998, 193)

In building his case for evaluating advantage in terms of what people are
able to be and to do (i.e. their capability), a central argument given by
Sen against welfarist alternatives is the problem of ‘adaptive
preferences’. He claims that subjective well-being alone is a poor guide
to real advantage, and thus a poor basis for interpersonal comparisons
of advantage, because many deprived people have had to come to terms
with their material and social deprivation in order to survive. Their
wants and aspirations, or their sense of well-being (happiness), cannot
be relied upon to track their authentic interests or even their physical
well-being since they may be a product of their circumstances. Sick
people may believe they are in good health; oppressed people may
express contentment about their treatment; and so on. As Sen puts it,

The deprived people tend to come to terms with their deprivation

because of the sheer necessity of survival, and they may, as a result,
lack the courage to demand any radical change, and may even adjust
their desires and expectations to what they unambitiously see as
feasible. It is thus important to take note of the fact that in the scale
of utilities, the deprivation of the persistently deprived may look
muffled and muted, but also to favour the creation of conditions in
which people have real opportunities of judging the kind of lives
they would like to lead. (Sen 1999a, 63)

This chapter is concerned with the problem of adaptive preferences.

First, what exactly are they? There are several definitions in the
capability literature (notably those by Jon Elster and Martha Nussbaum),
and Sen’s own use of the term does not identify a single mechanism but
rather a family of problems relating to deformations of practical
reasoning under conditions of material deprivation and social
oppression.
Second, what is the problem of adaptive preferences? It has been
argued, most notably by David Clark, that even very deprived people do
generally have aspirations for a better life. Therefore, he argues,
adaptive preferences are not a practical problem, but at most a merely
 89

theoretical issue (Clark 2009; Clark 2012). I argue that this dismissal is
premature. Adaptation presents a significant practical problem, not only
to people’s lives but also to the credibility of the capability approach to
evaluation.
Third, how can the capability approach address the issue of
adaptation? Following Sen’s own extensive use of Adam Smith’s concept
of the impartial spectator as “a device for reasoned self-scrutiny” (Sen
2012a, 104), I propose framing the problem of adaptation in terms of
preferences that would not survive its ‘transpositional’ scrutiny. This
makes two practical contributions. First, the impartial spectator
provides a model for judicious, respectful but engaged scrutiny of
suspected adaptation, which can come to reasonable judgements of
some cases, though not all. Second, with respect to remediation, it
suggests how “to favour the creation of conditions in which people have
real opportunities of judging the kind of lives they would like to lead”.
The impartial spectator framework conceptualises this in terms of
supporting the capability of individuals to become spectators on their
own lives, by giving them access to alternative epistemic positions from
which they can scrutinise their own values, desires and aspirations.
In the following section I outline Sen’s account of adaptive
preferences. Section II discusses the challenge posed by adaptation.
Section III presents Adam Smith’s impartial spectator. Sections IV and V
outline the practical contributions of this theoretical perspective.

I. WHAT ARE ADAPTIVE PREFERENCES?

A minimalist definition of ‘adaptive preferences’ is that they are
psychological states formed in reaction to personal circumstances.
There are at least three fuller accounts of adaptive preferences
discussed in the capability approach literature, and these are often
elided.91 Although this chapter is concerned with Sen’s account, I will
first outline the others, by Elster and Nussbaum, because each identifies
the problem and its solution in a different way.
Elster’s account of adaptive preferences in Sour Grapes (Elster 1982)
is the most precise, and therefore the most frequently referred to in the
capability literature despite the fact that it is designed for a quite
specific purpose unrelated to human development or the evaluation of

91
In the subjective well-being (‘happiness’) literature in applied psychology there is still
another definition of ‘hedonic adaptation’. For a comparison of the subjective well-
being and capability concepts see Teschl and Comim (Teschl and Comim 2005).
90

capability. Elster developed the concept as part of an attack on welfarist-

utilitarianism as the foundation for a theory of justice: “why should
individual want satisfaction be the criterion of justice and social choice
when individual wants themselves may be shaped by a process that
preempts the choice?” (Elster 1982, 219).
In order to make this critique work as an attack on the coherence of
utilitarianism, Elster defined adaptive preferences somewhat narrowly
and distinguished them from the effects of learning and experience,
wishful thinking, etc.92 What he identifies is a particular “sour grapes”
phenomenon in which someone begins by wanting something, then
realises that reaching it (the grapes) is impossible. The result is a
cognitive dissonance which is resolved by a non-cognitive (i.e. non-
conscious) re-evaluation of the situation and re-adjustment of her
preferences so that she genuinely believes that she no longer wants it.
To give a real world example, someone whose application to a
prestigious university is rejected may come to disvalue attending it and
genuinely prefer not to. The reason these preferences are problematic,
from the utilitarian perspective, is that they are entirely a causal
product of the feasible option set and are thus produced ‘behind one’s
back’. If the situation changes (the university accepts her after all) the
agent’s desire quickly reverts to what it had been before.
Elster’s concept of adaptive preferences is thus specific to a
particular purely procedural account of preference formation which
bypasses an agent’s autonomy. Preferences are adaptive when they are
induced in a particular unconscious way by changes in the feasible set
of options, whatever those preferences might be about. Elster
acknowledges that determining whether particular preferences are
autonomously formed is implausible, but notes that one can sometimes
prove that they aren’t the product of adaptive preference formation. If
one has the option to do or not do something, then one’s preference one
way or the other cannot be determined by one’s opportunity set.
Perhaps counter-intuitively, if one has an unfeasible preference (i.e. for
something which lies outside one’s opportunity set) then it also couldn’t
have been formed by this type of adaptation.

92
“[A]daptive preference formation has five distinctive features that enable us to locate
it on the map of the mind. It differs from learning in that it is reversible; from
precommitment in that it is an effect and not a cause of a restricted feasible set; from
manipulation in that it is endogenous; from character planning in that it is causal; and
from wishful thinking in that it concerns the evaluation rather than the perception of
the situation.” (Elster 1982, 226)
 91

Nussbaum defines adaptive preferences as a combination of

substantive mistakes and procedural wrongs (Nussbaum 2000, ch 2;
Nussbaum 2001). They are preferences counter to her anthropologically
derived list of central capabilities that have been formed in a non-
autonomous way “deformed by ignorance, malice, injustice, and blind
habit” (Nussbaum 2000, 114). With respect to justice, Nussbaum argues
that preferences counter to her list should be ignored for the purposes
of social choice.

A habituated preference not to have any one of the items on the list
(political liberties, literacy, equal political rights, or whatever) will
not count in the social choice function, and an equally habituated
preference to have such things will count. (Nussbaum 2000, 149)93

With respect to remediation, Nussbaum argues that the solution is

for a society to provide the capabilities on her list (leaving individuals
free to take them up or not). This is because Nussbaum considers that
adaptive preferences respond to learning and experience in the way
suggested by Mill in On Liberty (1859). People’s desires for the core
aspects of human flourishing have a unidirectional quality in that once
people have tried them they tend not to want to go back, or at least
appreciate the value of making this choice generally available to others
even if they themselves prefer not to make use of it (Nussbaum 2000,
153).
Sen’s own notion of adaptive preferences concerns the
internalisation of the circumstances of poverty and oppression in the
inner life of those concerned. The mental attitudes of such people may
misrepresent their objective wellbeing (“physical condition neglect”) and
also not reflect “the mental activity of valuing one kind of life rather
than another” (“valuation neglect”) (Sen 1985b, chap. 3). As a result,
their ‘utility’, whether construed as happiness or desire satisfaction,
does not track what they have reason to value.
Sen’s use of the term ‘adaptive preferences’ differs significantly
from that of Elster or Nussbaum (Qizilbash 2006; Qizilbash 2009). While
Elster limited his account to a single unconscious mechanism, Sen
suggests that adaptation can involve diverse mechanisms – from
hedonic adaptation, to informational deficiencies, to psychological

93
Nussbaum argues that this doesn’t make her account “totalitarian” because of the
various protections for pluralism and individual choice incorporated in her list
(Nussbaum 2000, 105).
92

temperament, to moral beliefs about justice and fairness – some of

which seem conscious. For example he notes that,

Deprived groups may be habituated to inequality, may be unaware

of possibilities of social change, may be hopeless about upliftment of
objective circumstances of misery, may be resigned to fate, and may
well be willing to accept the legitimacy of the established order. (Sen
1990b, 127 emphases added)

As a result Sen’s use of the concept of adaptation is both narrower

and broader than Elster’s analysis: Sen’s examples are narrower and
more specific (“landless labourer”; “subjugated housewife”) but its
scope is broader, incorporating, for example, ideological constructs
accepted by the oppressed (“objective illusions”).
Thus, like Nussbaum, Sen does not define adaptive preferences in
terms of any particular mechanism. But, unlike Nussbaum, he does not
define them by reference to a ‘correct’ substantive account of beings
and doings that people should universally value. Rather, the issue for
Sen seems to be the abnegation of agency involved in renouncing
aspirations for a better life in order to survive, and, at the extreme, a
denial of self. Thus, Sen notes that

It has often been observed that if a typical Indian rural woman was
asked about her personal ‘‘welfare’’, she would find the question
unintelligible and if she was able to reply, she might answer the
question in terms of her reading of the welfare of her family. (Sen
1990b, 126)

Adaptation thus strikes at the faculty for reasoning and self-

scrutiny, a faculty which Sen calls the “fourth aspect of the self” (Sen
2002e, 36) and considers constitutive to his conception of the person
(Sen 2002e, 46; Sen 1999d, 24).
Elster’s procedural account of adaptive preferences was driven by a
critique of utilitarian theories of justice for taking account “only of
wants as they are given” and to argue for the need for a “backward-
looking...analysis of the genesis of wants” (Elster 1982, 237 original
emphases). Nussbaum’s objective account was driven by a concern that
preferences against the availability of the items on her list should not be
counted in the social choice exercise. Sen’s concerns overlap with
Elster’s and Nussbaum’s, but he seems particularly focused on the
deprivation of agency implied by adaptation, in which a person’s
 93

capability to reflect on and scrutinise her values is restricted by

mechanisms which themselves reflect her circumstances of material and
social deprivation.
Yet, it has been argued that adaptive preferences are not a practical
problem. Or that if they are then they are not one the capability
approach is capable of identifying and dealing with. I therefore turn now
to discussing these critiques of the concept of adaptation.

II. TWO PROBLEMS OF ADAPTIVE PREFERENCES

The development economist David Clark has argued that there is little
evidence that adaptation is ubiquitous or significantly harmful (Clark
2009; Clark 2012). He is particularly critical of the paucity of systematic
empirical evidence to back up Sen’s or Nussbaum’s claims about
adaptation (Clark 2009, 29). He notes that his own fieldwork experience
suggests that,

while the poor and disadvantage[d] often report high levels of

happiness and life satisfaction (implying adaptation in terms of
subjective well-being), they are still capable of imagining,
articulating and demanding a substantially better or ‘good’ form of
life. (Clark 2009, 26)

In particular, Clark argues that the ability of many materially

deprived people to access and use information about alternative
lifestyles is much higher than appreciated. For example, in one survey
he reports, 90% of respondents living without access to medical facilities
nonetheless said they were “necessary for a person to ‘get by’” (Clark
2009, 32).94
In addition to his scepticism about the real world significance of
adaptation as a problem, Clark identifies a particular concern that its
logic produces a “retreat into social theory and high philosophy”,
instead of what he calls “empirical philosophy” - listening to the voices
of the poor who are the real experts about poverty (Clark 2009, 28-9). In
other words, the concept of adaptive preferences opens the door to
paternalism, to theoretical ‘experts’ determining in arbitrary ways what

94
Though, as Clark admits, this was not what the survey was designed for and its
results also revealed some evidence of Elster-type adaptive preferences. Apart from the
other 10% of people without medical facilities who thought they weren’t necessary,
two-thirds of people who lived in shacks thought shacks were sufficient to get by,
while only 25% of those who lived in houses or flats thought so (Clark and Qizilbash
2008, 536).
94

counts as an authentic aspiration. As Clark puts it, rather strongly,

The million-dollar question is why are so many development

ethicists, capability theorists and proponents of human need so
obsessed with the adaptation problem? In some cases at least, part
of the answer is that the adaptation argument can be used either
consciously or subconsciously to justify and privilege elitist
conceptions of well-being and development. (Clark 2009, 34–5)

The issues Clark raises present an important challenge to the

concept of adaptive preferences. The argument about paternalism in
particular has been echoed by philosophers, most systematically by the
philosopher-economist Robert Sugden (Sugden 2006). Sugden raises the
under-examined role and possible paternalism of “the moral observer”
in Sen’s account of evaluation: “The idea that ‘we’, as ethical theorists,
can claim to know better than some particular individual what is good
for her seems to open the door to restrictions on freedom (Sugden 2006,
34).”95 So it is worth responding to the twofold challenge: what is the
practical significance of adaptation and can the capability approach
address it without paternalism?
As to the first challenge, Sen in no way suggests that adaptation is
ubiquitous and notes explicitly that “that special problem .... does not,
of course, yield a general case for believing .... that ‘ethical theorists can
claim to know better than some particular individual what is good for
her’” (Sen 2006d, 88). Yet adaptation certainly does occur (as Clark
acknowledges), and is hardly a new discovery (it appears in the work of
Adam Smith, Mary Wollstonecraft, Karl Marx, and J. S. Mill, among
others).
Sen also argues that adaptation is not only a deprivation in itself (of
the freedom to reflect on and choose one’s goals and values) but can
also be a real problem for objective well-being in two ways. First, it can
take the form of mistaken perceptions of the way the world works that
have a significant detrimental effect on a person’s well-being, since, for
example, someone who doesn’t realise they are sick is unlikely to seek
medical treatment.96 Second, adaptation also concerns perceptions of
how the social world works, which can induce people to see oppressive

95
Sugden seems to have in mind Sen’s use of the ‘device’ of Adam Smith’s impartial
spectator (Sen 2006d, 88), and this provides a further motivation for my exploration
and elaboration of that device here.
96
As Sen noted in his comparison of self-reported health and actual life expectancy in
different parts of India (Sen 1993a, 134–5).
 95

conditions as legitimate, fair and entirely reasonable. They thereby

influence people’s sense of what kind of life they deserve to have, and
also what kind of life other people think they deserve to have. These
problems are connected, since social norms may often be understood as
facts about the natural world. For example, perceptions about what girl
and boy children need can influence perceptions of what they deserve to
get (as Sen’s empirical research on children’s health has shown).97 On the
assumption that the well-being of one’s children is something one has
reason to value, if a child’s health suffers because her caring parents
hold such mistaken beliefs, then the parents’ lives too would seem to be
harmed.
That example is pertinent, since Sen’s most extensive and systematic
analysis of adaptive preferences has concerned issues of gender
inequality. Indeed, he notes that it was his exposure to feminist ideas
that stimulated his interest in adaptation mechanisms (Sen 2003a, 319–
324). It is rather difficult to dispute the scale and urgency of the real
world problems for well-being that can be related to invidious gender
norms (held both by men and women). Sen points out that the relative
deprivation of women in a country can often increase even while
aggregate well-being increases (in terms of longevity, literacy,
nourishment, etc.). He has also shown that up to 100 million women are
missing from the world’s population as a result of sex-selective abortion
and physical neglect relative to males at every stage of life (Sen 1990d).98
In particular, Sen has contributed an original approach to
understanding the relationship between men and women within a
household in terms of ‘co-operative conflicts’ which builds explicitly on
the problem of adaptation (Sen 1984a; Sen 1989c; Sen 1990b). This
framework unpacks ‘the household’ into its constitutive persons with
their own distinct interests, and perceptions of their interests (usually
amalgamated together or ignored in standard micro-economic
representations). Men and women within a household are considered as
engaged in a ‘bargaining’ exercise over the division of the gains of co-
operation. That division will be influenced by relative bargaining power,

97
Girl children were perceived to have different (lesser) care needs than boy children.
For example, girls had to be much sicker to be taken to the hospital. But this was not
understood by the parents as gender discrimination. Unsurprisingly, however, the
result was that boys were generally better nourished and in better health. (Kynch and
Sen 1983; Sen and Sengupta 1983)
98
These are women who would be expected to be alive given the number of men who
are alive, if natural (biological) birth ratios and survival rates under conditions of equal
nutrition and care were determining the outcome.
96

which may be influenced by men’s generally physical dominance and

better relative ‘fall-back’ positions (i.e. their ability to prosper outside
the marriage). But it may also be significantly influenced by perceptions
of fairness concerning both the status of different parties’ contributions
(for example by placing a higher value on household contributions from
non-domestic waged labour), and also the perception of legitimate
needs.
This is a problem for justice, but is it a problem for the evaluation of
capability? I believe it is, for two reasons. First, because the evaluation
of individual advantage “require[s] some understanding of what the
person would want to have and have reason to value having” (Sen 2002e,
5). Adaptation makes this difficult to discern. Second, because such an
evaluation should be able to note the specific deprivation involved in
adaptation: whether people are in the grip of “a valuational mist that
engulfs all and which works by making allies out of the victims” (Sen
2003a, 320).
Consider for example the introduction of cheap prenatal gender
screening across India and China (closely connected with a birth ratio
dramatically skewed towards boys in some regions). Should we evaluate
this new ability of even quite poor women to choose the sex of their
children as enhancing their capability? While being able to guarantee not
having a female child might be quite a rational thing to want in
particular social circumstances, that is not the same thing as being what
the women themselves have reason to want. Rationality in this sense is
concerned with a merely instrumental aspect of practical reason -
relating means to ends - rather than the determination of ends
themselves. The rationality of preferences thus does not demonstrate
their reasonableness (Sen 2002e, 39). If such a preference is the result of
rational acquiescence to invidious social norms and the coercive
demands of family members, then it is a textbook case of adaptation.
Thus the fundamental problem posed by adaptive preferences
relates not to their ubiquity, but to the uncertainty they introduce to the
evaluation of advantage in terms of what the people concerned have
reason to value. Without evidence of the reasoning process, “The mental
metric of pleasure or desire is just too malleable to be a firm guide to
deprivation and disadvantage” (Sen 1999a, 63).
Let me now turn to the second challenge, can the capability
approach address the problem of adaptation without paternalism? One
obvious but limited route is to distinguish what Sen calls the “well-being
 97

perspective” from the “agency perspective” (Sen 1985a; Sen 1993b).

Whereas welfarist-utilitarians only consider subjective well-being, the
capability approach can directly identify failures and successes in
objective dimensions such as literacy, nourishment, health, and so on.
Whether or not someone is under-nourished is a factual matter that can
be determined in an empirically objective way. If one accepts Sen’s
argument that such facts matter to the evaluation of well-being then this
is an important exercise that can be carried out even where adaptation
is significant. For example, the well-being of the “typical Indian rural
woman” who cannot distinguish her own well-being from that of her
family could be evaluated in terms of health, nutritional status, literacy
and so on (Sen 1990b, 126). Thus, if the capability approach were
restricted to considering objective well-being achievements, and to
evaluating this in terms of a relatively narrow set of dimensions
generally agreed to be essential, it would be able to deal with the
problem of adaptive preferences by parenthesising it (together with
subjective well-being in general).
Yet the cost of this narrow use of the capability approach is that its
foundational concern with agency and freedom would be excluded from
practical application. While it seems reasonable to suppose that health,
literacy, nourishment, and so forth are capabilities everyone has reason
to value, it is not reasonable to suppose that they are the only things
people have reason to value, or even the things that people have most
reason to value. The objective well-being use of the capability approach
would be unable to distinguish between fasting and starving, for
example, and would evaluate Gandhi’s use of fasting as a powerful tool
of political and religious protest only in terms of its negative effects on
his physical well-being.
In Well-being and Agency (Sen 1985a), Sen notes that consideration
of well-being is in terms of externally determined criteria, such as
prevailing standards, while consideration of agency takes the person
concerned as the source of the relevant evaluative criteria. To consider
someone’s advantage only from the perspective of well-being is, Sen
suggests, appropriate where agency is absent, that is, for young children
and the severely cognitively disabled (Sen 1985a, 204). The risk of
paternalism here is eloquently expressed in Lionel Trilling’s warning
about the wilful, imperious and impatient nature of our moral passions,
98

[W]e must be aware of the dangers which lie in our most generous
wishes. Some paradox of our natures leads us, when once we have
made our fellowmen the objects of our enlightened interest to go on
to make them the objects of our pity, then of our wisdom, ultimately
of our coercion. It is to prevent this corruption, the most ironic and
tragic that man knows, that we stand in need of the moral realism
which is the product of the free play of the moral imagination.
(Trilling 1948, 27)

Trilling points directly to the temptations towards despotic

paternalism inherent in benevolence. That temptation seems especially
likely in the case of development work, where asymmetries of power,
and social and cultural distance are especially great. Trilling was a
literary critic and, like Nussbaum (e.g. Nussbaum 1995; Nussbaum
1997), saw literature as an important means for learning to consider
other people’s lives and how their lives feel to them. In the argument
that follows I will show that we need more than empathy with others,
and that Adam Smith’s impartial spectator provides a more general
account of the importance and role of “the free play of the moral
imagination”.
This section has considered whether adaptive preferences are a real
problem, and, if they are, whether the capability approach itself is in a
position to address them without paternalism. I have argued that the
phenomenon of adaptation can indeed be a big problem for people’s
capability to live the kind of life they have reason to value. Their harm is
both instrumental (negatively affecting other valuable capabilities) and
intrinsic (a deprivation of what Sen calls “the fourth aspect of the self”).
I have also argued that the capability approach’s main way of dealing
with adaptation, by setting all subjective values aside, is an evasion of
the central challenge they pose: to adequately consider how advantaged
a person is from the agency perspective i.e. without falling back entirely
on the limited informational base of how well they are doing in terms of
what are generally considered (by other people) to be valuable
functionings. That evasion limits the scope of capability evaluation, and,
moreover, can have a paternalistic character insofar as it treats
responsible agents as if they are children or severely cognitively
disabled.
To the extent that the agency perspective is considered constitutive
of the capability approach, it requires an explicit evaluative framework
for dealing with adaptive preferences that is sensitive to the
deformation in practical reasoning that can be induced by deprivation,
 99

but also liberal in its fundamental respect for the agency of those
concerned. It should address Clark’s concerns about retreating into high
theory and neglecting the voices of the poor. That framework, I suggest,
can be constructed from a suitably adapted model of Adam Smith’s
impartial spectator.

III. ADAM SMITH’S IMPARTIAL SPECTATOR

In the Idea of Justice, and elsewhere, Sen makes extensive use of Smith’s
impartial spectator in thinking about the demands of justice and in
particular for thinking about their reach across cultural and political
borders, for what he calls ‘open impartiality’. The basic idea, as Sen puts
it, is “the requirement when judging one’s own conduct, to ‘examine it
as we imagine any other fair and impartial spectator would’” (Sen 2002d,
449). I won’t be concerned here with Sen’s use of the impartial spectator
for cosmopolitan evaluation, however, but will focus instead on whether
the impartial spectator can play a different theoretical and practical
role, in the conceptualisation, evaluation, and remediation of adaptive
preferences. In this section I will provide an abbreviated outline of
Smith’s idea of the impartial spectator and its underlying moral
psychology from his Theory of Moral Sentiments (Smith 1759) focused
on those aspects that are relevant to the specific issue of adaptation.99
Smith argued that moral life was structured around an emotional
economy mediated by sympathy and oriented towards harmony. In the
sentimentalist tradition of the Scottish Enlightenment, Smith
understood emotions as having cognitive and normative content,
incorporating actors’ general beliefs and concepts and their specific
judgements of particular situations. That cognitive content could be
vicariously grasped and evaluated by a critically engaged spectator who
conducts three distinct operations. First, diagnosing the emotions
expressed by an actor (from their expressions, talk, and actions).
Second, considers how she would feel in the actor’s position (sympathy).
Third, bringing the results of those together and evaluating whether the

99
For a fuller critical analysis of Smith’s impartial spectator and its role in his moral
philosophy see D. D. Raphael (Raphael 2007).
Note that Nussbaum has also made some use of Smith’s account in Poetic Justice in
support of her “analogical reasoning” approach to the philosophy of law (Nussbaum
1995 particularly 72-77). I do not believe she has Smith quite right, however, and will
be pursuing a different interpretation. (She appears to reduce Smith’s account of
sympathy to the exercise of empathy and the feeling of pity, and refers to his impartial
spectator as the “judicious spectator”.)
100

actor’s emotions are more or less appropriate for the situation as she
understands it.
In this account sympathy should be understood adverbially, as
applying to an emotion felt sympathetically, according to the spectator’s
own feelings when imaginatively transposing herself into an actor’s
position. Sympathy is thus not necessarily a concordance of feelings
between actor and spectator because their assessments of the relevant
situation may vary. It is the situation that generates the spectator’s own
sympathetic feelings, which she then uses to evaluate the propriety of
the actor’s behaviour. For example, if one saw someone being mistreated
and they said ‘thank you’ afterwards, one would have the strong first
impression that there was something wrong with the victim as well as
with the assailant. Smith’s concept of sympathy thus refers to a rigorous
procedure of transpositional evaluation, rather than an emotional state
of unquestioning compassion (as the word ‘sympathy’ is now often
understood). Thus,

The propriety of a person’s behaviour depends not upon its

suitableness to any one circumstance of his situation, but to all the
circumstances which, when we bring his case home to ourselves, we
feel should naturally call upon his attention. If he appears to be so
much occupied by any one of them as entirely to neglect the rest, we
disapprove of his conduct as something we cannot entirely go along
with, because not properly adjusted to all the circumstances of his
situation. (Smith 1759, V.ii)

Sympathy depends on what Smith considered a natural human

capacity, as fundamentally social creatures, for imagining ourselves in
another’s situation, combined with a natural human disposition to seek
harmony (also seen in our propensity to “truck, barter, and exchange”).
It thus works through and over our Protagorian limitations: that we can
only judge others by our own standards (Smith 1759, I.I.29). Successfully
imagining oneself in another’s situation is an exercise in interpretation.
It thus takes place against a background understanding of human
nature and morality. This is not a view from nowhere, but rather reflects
the spectator’s own expectations and standards. Individuals naturally
develop these over their lives, in their reflective judgements about their
socially learned norms and ideas, combined with their own personal
experiences (which can include vicarious experiences such as provided
by news reports and the fictional depictions of the human condition
found in drama and literature) (Fleischacker 1999, 41–55).
 101

Smith sees this dialectical relationship literally at work in the moral

development of individuals and societies. Children, for example, quickly
learn to view themselves as other people do and thereby to internalise
the social standards they learn from what brings praise or blame. In
turn, those standards are of great social interest and are continually
discussed, analysed, and debated even if more generally in the form of
moralised gossip than high philosophy (Fleischacker 1999, 51).
By internalising this sense of social propriety we learn to see
ourselves from the perspective of other people in general, or to imagine
how we would appear to any disinterested bystander or impartial
spectator (Smith 1759, III.iii). Yet Smith did not believe that the
mechanism of the impartial spectator stopped there. Once learned, the
skill of imagining an impartial spectator can be logically extended as a
resource for Socratic self-scrutiny. One can imagine an ideal impartial
spectator who is the distilled essence of spectatorship: engaged but
detached; without biases, but with all the knowledge about the situation
one possesses.100 One can then ask, can this impartial spectator
sympathise with these values or actions of myself or others?
Smith argues that this device allows the ethically developing agent to
transform his original sense of social propriety into a sense of morality
proper. In this mode the ethical agent can turn his scrutiny on the
original social norms themselves. Not only does the ideal impartial
spectator judge one’s own behaviour in the light of the standards one
acknowledges in one’s calmer moments one should uphold, whether of
honesty, prudence, or proper manners. It also critically evaluates the
quality of those standards and one’s understanding of them.
Smith emphasised that this use of the impartial spectator required
wisdom as well as motivation (virtue). Arriving at and properly
understanding the best moral standards requires careful thought rather
than social osmosis, in terms, for example, of the examination of the
qualities and external coherence of one’s current moral beliefs. For
example one can test the basis of the judgements one makes of others
by applying them to oneself. This can both expose mere prejudices and
prompt us to revise our moral norms and treatment of others.101 Smith

100
It should be noted, however, that Smith’s impartial spectator remains firmly
grounded in the limited capacities and information of actual people. It is thus ‘ideal’
only in aspiration, and is not a contribution to ideal observer theory, as Rawls for
example interprets it (Rawls 1999, §30).
101
Smith, for example, used the contemporary European condemnation of Chinese
foot-binding practises as a mirror to reveal the parallel injustice of the unexamined
102

suggests that this capacity allows the emergence of morality proper,

rather than merely the reflection of dominant social norms. Indeed it
provides the resources and authority for an individual to question and
dissent from prevailing social norms.
Smith attempts to turn the weakness of the parochial beginnings of
our moral practises into a strength, by showing how the impartial
spectator works as a ratchet to generate support for an objective
transpositional ethical perspective out of the base materials of our
ordinary prejudices. Yet the idea of the impartial spectator is normative
as well as descriptive – it should also improve our performance as moral
agents. For example, it should influence how we assess others, so that
we try to sympathise with others’ situations as an ideal impartial
spectator would, setting aside our prejudices but not our knowledge.
This both provides a stronger motivation to be open minded in one’s
moral evaluations of other people’s situations and conduct, and holds
those evaluations to a high standard of objectivity.102
Smith’s system uses the force of socially shared norms to get up and
running but the system itself contains checks against both the risks of
narrow minded parochialism and also against a flawed or arrogant ideal
impartial spectator. This is necessary because, while such ethical
reflection may aim for the transcendental impartiality of the ideal
spectator, it can never entirely achieve this since it is, in the end,
founded on the same social tribunal upon which it seeks to judge (Smith
1759, III.ii). The ideal impartial spectator is thus required to toggle
between our social and ethical understandings to refine our common
sense prephilosophical concepts and practices (Griswold 1999, 24),
rather than providing omniscient and final judgement from on high. It is
a humble creation, well aware of its own fallibilism and therefore ready
to reconsider its judgements and the general moral rules it lays down
for itself, in the face of any relevant new evidence, including strongly
expressed or well-reasoned opposition from other people.
Smith’s moral psychology offers a framework for understanding
what is at stake in adaptation – the capability to be a spectator upon
oneself – and the nature of adaptive preferences - values, preferences,
and aspirations which the impartial spectator would not go along with.
In the next sections I will show how this perspective can provide

European custom that compelled women to wear physically disfiguring corsets (Smith
1759, V.i.8).
102
And this is something Smith illustrated in his own practise, in considering non-
European cultural mores (Pitts 2005, chap. 2).
 103

practical guidance for the methodology of evaluation and the

remediation of adaptation.

III. EVALUATION
The capability approach’s concern with the effective freedom of
individuals to lead a life they have reason to value requires a
counterfactual mode of evaluation. The observer must assess not only
those functionings an agent achieves but also what other valuable
opportunities they really had. This is further complicated in that the
source of valuation is supposed ultimately to be the individuals
concerned (the life they have reason to value), and so the observer must
not only come to a reasonable assessment of an actor’s situation, but
also consider the actor’s own evaluation of that situation and its
reasonableness in turn. As Sen notes,

[A]n assessment of the opportunities a person has would require

some understanding of what the person would want to have and
have reason to value having…..[but] the concept of rationality must
accommodate the diversity of reasons that may sensibly motivate
choice. (Sen 2002e, 5)

Close attention to the agent’s particular case is required and one

cannot read off an agent’s lack of freedom directly from her well-being,
nor her rationality from the well-being implications of her choices. It is
part of an agent’s freedom, for example, that she may choose to exclude
certain options which might enhance her well-being but which are
excluded by social or deontological norms that she chooses to obey (Sen
2002e, 18). Yet this freedom also implies that she cannot be required to
be moral, in the sense of adhering to the ‘correct’ or indeed any moral
theory or precept. As Sen puts it, “The demand of rationality is not so
much to require conformity to any particular set of goals or values, but
to demand that both one’s goals and non-goal values should be
supportable through careful assessment and scrutiny” (Sen 2002e, 41)
The concept of the impartial spectator makes two contributions to
the methodology of evaluation, concerning the ethos of what Sugden
calls the “moral observer” and guidance for diagnosing adaptation. As I
noted in outlining the impartial spectator, Smith believed that his
concept performed real normative as well as descriptive work, in
holding people to a higher standard not only in their self-regarding
actions and thoughts, but also in their understanding and normative
104

judgements of others. The dichotomy between actor and spectator can

be applied to the action of spectating itself, so that the person acting in
the role of ‘moral observer’ - charged with judging the effective freedom
of strangers - is called on to consider her conduct and methodological
choices with regard to what an impartial spectator can go along with.
The key elements of this seem to me to include a presumption that
those one is assessing are indeed responsible agents with reasons of
their own; a commitment to understanding their situation as the agents
concerned do; and a commitment to continual scrutiny of one’s own
expectations and normative standards.
Although the observer may ultimately disagree with the actor’s own
assessment, her first imperative is to try to find a way to agree, and
therefore she will first attempt to bring the actor’s case home to herself
by imagining herself as far as possible in the actor’s situation, while
retaining her own perspective - identity, standards, and knowledge. That
knowledge should include the best possible synthesis of the findings of
relevant empirical research, and of the possibility of alternative social
arrangements. Alexander Broadie suggests that the relationship is
characterised by the same sequence of “surprise, wonder, and
admiration” found in Smith’s writing on science (and structurally
parallel to the famous sequence of “truck, barter, and exchange” found
in The Wealth of Nations) (Broadie 2006, 177). First comes surprise, in
which the moral observer identifies a dissonance in his own and the
actor’s reactions to or judgements of a situation. This sets the dutiful
observer to wonder at the actor’s behaviour by subjecting the case to
further scrutiny including putting himself in the actor’s shoes. Finally, if
the exercise is successful, the spectator will resolve the dissonance and
come to admire, i.e. go along with, the actor’s behaviour.
The spectator begins from the assumption that the actor is a
perfectly competent agent, but may be struck by aspects of the actor’s
situation that conflict with her own presumptions and raise a question
over the freedom or agency of the actor (surprise). Surprise is most
likely when a person’s situation and self-assessment appear extremely
out of line to the spectator’s normal presumptions, for example where
very basic capabilities such as freedom from violence are absent but the
actor not only expresses satisfaction with that state of affairs but
endorses it.
Surprise induces the spectator to attempt to understand this
disjunction by considering, as an impartial spectator should, whether he
 105

has missed or misunderstood any significant elements of the situation

that might make the actor’s behaviour intelligible and reasonable
(wonder). This can include consideration of relevant empirical work, for
example within sociology and psychology, which identify mechanisms of
adaptation, and their symptoms, for example, in the case of women
unable to consider leaving a violent relationship. If the dissonance
continues it may become reasonable to consider whether the actor
herself has really considered (or had the chance to consider) her
situation properly. For example, it may seem that she has failed to
consider all the salient aspects of the situation, or is ignorant of
important facts about the issue or alternative perspectives on it. In
seeking a reasonable explanation for this behaviour, the spectator may
seek more information and, if feasible, even ask the actor to provide
reasons to go along with her behaviour. This can continue until either
the spectator-actor evaluations converge (admiration) or (further)
convergence appears impossible to reach.
In this process, the spectator may be forced to reconsider some of
his original presumptive interpretative rules as perhaps not applicable
to this case in the way she expected or perhaps discard them as
prejudices. Close scrutiny of the case of a transgender person who
considers that she is a woman trapped in a man’s body, for example,
may lead the spectator to revise his preconception that such feelings are
symptoms of mental illness, and thus substantially alter his evaluation
of that person’s advantage in terms of her capability to live the kind of
life she has reason to value.
In another case, such as of the preference for gender-selective
abortion in some parts of India and China, close scrutiny of the relevant
sociological literature and qualitative interview surveys may lead the
spectator to confirm his initial impression that this is very likely
induced by oppressive circumstances. Rather than being considered as a
freedom that these women have reason to value, access to sex-selective
abortion would then instead better be seen as an aspect of the
perpetuation of unfreedoms associated with the general social
disvaluation of women. And this would be consistent with the
conclusion that, in their social and domestic circumstances, access to
sex-selective abortion significantly enhances those women’s well-being.
That is to say, such a conclusion suggests that if these women were able
to make a choice about the kind of world they lived in, they would
prefer one that did not require them to settle for such aspirations.
106

There are obviously major limitations to this approach. Judging

another person’s use of, or capability for, self-reasoning is inherently
difficult and only relatively clear cases may be judicable by an external
impartial spectator. As Sen notes, “Seeing rationality as a discipline of
reasoned scrutiny does not yield – or even attempt to provide – any
sure-fire ‘test’ of whether a person is or is not being rational” (Sen
2002e, 48). For example, sometimes there may be good reason to
conclude that some adaptation is occurring - such as evidence of the
general acceptance of iniquitous gender norms about employment
aspirations and treatment - without being able to say whether any
particular individual’s values and aspirations about work are the result
of adaptation, rather than a personal conception of the good which
incorporates social norms in a way that makes them their own.

V. REMEDIATION

An Afghan girl today, kept out of school and away from knowledge
of the outside world, may indeed not be able to reason freely. But
that does not establish an inability to reason, only a lack of
opportunity to do so. (Sen 1999d, 26)

As I noted, Sen argues that the capacity for “self-reasoning”, of

submitting one’s own goals and values to critical scrutiny, is a
constitutive feature of human persons. Yet that innate capacity requires
the right kind of opportunities to be realised. Nussbaum considers that
once universal access to the objectively valuable functionings she
identified could be established, maladaptive preferences not to have
those capabilities (formed through ignorance, malice, and so forth)
would wither away. For example, one can provide services and security
for people trying to escape domestic violence without determining who
is or who is not suffering from it. And one may expect that many people
will manage to seize such opportunities, as in the case of Vasanti (one of
Nussbaum’s interviewees). However, as Qizilbash notes, in such cases
one should not count the provision of such opportunities as enhancing
people’s capability sets, since one cannot tell whether they are actually
able to choose to take them up (Qizilbash 2008, 61–2).
In contrast, Sen emphasises a different kind of opportunity – the
opportunity to reason about one’s preferences.103 Though these two

103
“What the critics of unreasoning acceptance of persistent deprivation want is more
reasoning about what ails the perennial underdogs, with the expectation that, with
 107

approaches are not mutually exclusive, their specific policy

recommendations are quite different in their focus. The impartial
spectator framing of the problem of adaptation elaborates Sen’s general
concern with opportunities for reasoning into a more specific
programme of giving people the opportunity to become spectators on
their own lives by giving them access to alternative epistemic positions
from which they can scrutinise their values and aspirations.
Sen is a strong proponent of the contribution of “transpositional
scrutiny” to overcoming the problems of parochial ignorance or
prejudice. Different perspectives are treated as epistemic resources
rather than separate islands. By familiarising oneself with a number of
such perspectives one can come to synthesis a much better overall
picture of the object one is looking at than any one perspective could
supply, and also identify the inadequacies of some perspectives (cf
Anderson 2003).
Although Sen has not explicitly referred to Smith’s impartial
spectator in the context of adaptive preferences, he has argued that it
facilitates the transpositional approach to evaluation he takes to be
central to practical reasoning and human agency.

[T]here is a crucial need (1) to examine the world in the light of the
values we have, and (2) to examine these values themselves and to
scrutinize the justification we can give for them. It is for facilitating
a serious and deep re-examination that Adam Smith invoked his
innovative device of an imagined ‘impartial spectator’ who looks at
our choices and behaviour, using a wide set of alternative values –
possibly quite different from those which we instinctively accept –
and invites us to think critically about our own values, in the light of
other ones which we can consider and scrutinize. (Sen 2012a, 103)

Sen’s concern with transpositional scrutiny is particularly clear in

Positional Objectivity (Sen 1993a) which analyses the significance of an
observer’s position for objective assessment. Sen notes the phenomenon
of “objective illusions”, such as that the ‘moon is the same size as the
sun’ which are strictly false (hence ‘illusions’). Nevertheless they are
objective illusions in the sense that anyone else placed in the same
‘position’ - i.e. with the same sensory and information parameters
(including related knowledge about optics and astronomy) - would see

more scrutiny, the ‘well-adapted’ deprived would see – and ‘feel’ – reason enough to
grumble.” (Sen 2009a, 275)
108

things the same way. They are not delusions. They are position-
dependent but person-invariant.
Interestingly, Sen argues that this analysis is particularly relevant to
diagnosing and countering adaptation (Sen 1993a, 134–5). Sen sets out
the difference between self-assessed morbidity between the two Indian
states of Kerala (high) and Uttar Pradesh (low). If self-assessed morbidity
levels reflected actual sickness levels then we would expect the people
in Uttar Pradesh to be generally healthier, but the data on life
expectancy suggest the opposite (Kerala: high life expectancy; Uttar
Pradesh: low). There is a dissonance here that needs to be resolved, and
Sen suggests it is due to the fact that the people in Uttar Pradesh have,
on average, different understandings and expectations of health, and
that this is related to their much lower levels of public health provision
and education, especially about illness prevention (as compared with
Kerala). The self-assessed health statistics in Uttar Pradesh may be
rational from the restricted perspective most people there occupy
(without reasonable information and expectations about health), but
they do not stand up to transpositional scrutiny. The solution here
seems obvious – to introduce people to other epistemic positions on
health, most obviously by increasing health provision and health
education in the area, so that the “objective illusion” of good health
disappears together with the position from which it made sense.
When Sen turns to gender disparities in self-assessed well-being, he
notes that the women’s perspective here is entangled in opaque cultural
norms which they themselves may uphold but which are rarely made
explicit because they are not the kind of thing one talks about (Sen
1993a, 135–6). Nevertheless their perspective could be represented as a
very specific set of positional parameters, and, Sen implies, the exercise
would be a very useful one because it would make implicit tacit norms
and beliefs explicit, and this is a first requirement for being able to
subject them to critical scrutiny. It invites those concerned to be
spectators on their own lives.
Providing people with opportunities to become spectators on their
own lives can be justified as a development goal in its own right and as
instrumentally significant for other valuable capabilities. It can take
many innovative forms, from dialectical interventions by development
agencies to entertainment education.104 For example, in one action

104
A further important opportunity, discussed further in the following chapters, is
political participation, which challenges participants to articulate their values explicitly
 109

research programme focused on eliciting and supporting women’s life

aspirations, the researchers included the question,

Imagine that you have a birthday or a celebration many years from

now. Imagine you are now older. Imagine you have had a good life, a
happy life. Your family and the community have come together to
celebrate your birthday and your achievements. Imagine that you
also feel very happy and thankful that things have worked out so
well for you. What is it that you have achieved in this life you have
imagined? What is it that you see in your life when you look back?
(Conradie and Robeyns Forthcoming, 9)

Other examples include the use of character driven narratives, rather

than information or exhortation driven lectures, in public education
radio and television broadcasts. This is a more demotic and carefully
targeted version of Nussbaum’s promotion of literature for
understanding other lives. The aim, ultimately, is to provide people with
the opportunity to view their own values, desires, and aspirations from
a critical distance by giving them imaginative access to other possible
lives, as alternative epistemic positions. A famous example of this is
Taru, a radio soap opera series broadcast by All India Radio (AIR) in
Bihar state during 2002/3 (Singhal 2010; “Soap Operas - Art Imitating
Life” 2012). One of the story lines was of a little girl who demanded to
have a birthday party like her brothers got to have (unlike girls, boys in
Bihar generally have their birthdays celebrated as well as receiving
better nutrition, education, care, and more positive attention from
family and community adults). The story line built up over a number of
weeks, following the persistent birthday girl through the tribulations
and joys of organising her party, from sending out invitation cards to
choosing her cake. Soon after it was broadcast, birthday parties for girls
were being reported all over the state, along with a great deal of explicit
discussion about gender equality.

and convincingly (i.e. in a way that others can go along with) and which also submits
value claims to the public scrutiny of various perspectives.
110

VI. CONCLUSION

The underlying motivation [of the capability approach] – the

focusing on freedom – is well captured by Marx’s claim that what we
need is ‘replacing the domination of circumstances and chance over
individuals by the domination of individuals over chance and
circumstances’. (Sen 1989a, 44)

Adaptation to material poverty or social oppression seems to

represent the quintessential dominance of social and material
circumstances over people’s lives, and is a significant harm and a
deprivation in its own right. Diagnosing and addressing it must be a
particular priority for the capability approach.
The problem is in doing so without making the mistake either of
treating agents as children or of endorsing, and thus perhaps
entrenching, the oppressive circumstances. I have argued that Adam
Smith’s device of the impartial spectator can play a role in diagnosing
and remediating (or preventing) adaptation. First, those concerned with
the evaluation of the capability of individuals should adopt the model of
the impartial spectator in their ethos and methodology. This is not the
assertion of a single ‘reasonable’ substantive perspective (such as
Nussbaum’s list of central capabilities), but a concern with tracing the
reasoning of the actors involved. Second, from this perspective the
problem of adaptation lies in the inability of individuals to consider
their lives – and their values, desires, and aspirations – from a critical
distance. Addressing this requires providing people with opportunities
to become spectators on their own lives.
 111

Chapter 5: Transformation Without Paternalism105

So act as to treat humanity, whether in thine own person or in

that of any other, in every case as an end withal, never as means
only.106

Human development is a transformational project.107 It is about

changing people’s lives for the better. This is so obvious that one might
suppose it wouldn’t need saying. And yet the transformative nature of
development has controversial implications that are rarely explicitly
considered. Do the people concerned agree that their lives are deficient?
Do they freely consent to the process of development, or is it something
that is done to them? Are they better off after development than they
were before, or merely changed? Intuitively, something is amiss if we
cannot give positive answers to these questions. Should we even call it
development? This chapter is concerned with elucidating the ethical
implications and constraints that follow from examining the idea of
development as transformation. These issues have, I believe, been
relatively neglected in the development literature, including that of the
human development approach.
Other ‘conventional’ accounts of development may fudge the issue
of transformation by speaking only of enhancing people’s capacities to
live well, for example in terms of meeting people’s existing basic needs,
or increasing their ability to satisfy their existing wants (perhaps by
increasing their ‘budget’ through economic development). In such
accounts, people are understood as essentially passive and static with
respect to the development process itself, and the challenge of
development is merely to help them to better live the lives they now
have. In contrast, the human development approach has directly

105
This chapter is adapted, with minor amendments and the inclusion of some new
material (the communitarian critique and the Gram Vikas case study), from a paper co-
written with John B. Davis. I am grateful for his collaboration and for his permission to
re-use our joint work here.
106
This quotation, from Immanual Kant’s Fundamental Principles of the Metaphysics of
Morals, is the epigraph with which Sen begins his overview of the capability approach
Development as Capability Expansion (Sen 1989a, 41).
107
The ‘human development approach’ is the development focused interpretation of
Amartya Sen’s capability approach promoted by the United Nations Development
Programme (UNDP). In this chapter ‘capability approach’ and ‘human development
approach’ are used synonymously.
112

transformational goals: to expand people’s freedom to live the kinds of

lives they have reason to value. Not only is its focus on the people rather
than the means of development, but it actively engages those concerned
in two ways in an approach to development understood “as a
participatory and dynamic process” (UNDP 1990, 11). Its goal is that
people become more truly the authors of their own lives, and it
considers this active authorship also the means by which development
should be achieved (cf. Sen 1999a, 18–9). The human development
approach thus has no choice but to face up to the conceptual and
ethical implications of personal transformation.
The general risk in neglecting the ethics of transformation is of
paternalism, of substituting one’s own values for those one is trying to
help. This is an ever present danger in work on development and one
which can creep in all too easily in the company of good intentions. This
chapter seeks to make three distinct contributions to preventing
paternalism in development, one theoretical and two practical.
First, although the capability approach understands persons as
agents, it represents people in terms of their location in the capability
space (that is, in terms of the set of functioning n-tuples to which they
have effective access). This way of representing persons can be
problematic to the extent that it suggests a thin, static, and passive
sense of personal identity that distorts evaluation or policy. I address
this by introducing the concept of a ‘personal identity capability’, which
is an account of personal autonomy analyzed as a capability. Thus
theorized, persons can be understood not only in the general, somewhat
diffuse sense of the agent as a “doer and a judge” (Sen 1985a, 208), but
considered more precisely in terms of their capability to take up a
reflective stance towards themselves as a person persisting and yet
developing over time, and to make plans and choices accordingly.
Second, this representation provides a new goal and side-constraint
for development practice. The protection and enhancement of
individuals’ personal identity capability should be recognised as an
important goal in its own right. And it should also be recognised as a
requirement for any policy to count as truly human development.
Applying this is not straightforward, however, and this leads me to
make two specific proposals for development practice. I first introduce
the principle of ‘free prior informed consent’ as a central ethical concept
for organizing and guiding ethical reflection in the practice of
development. Second, I argue that ‘democratic development’, in which
 113

the people concerned participate in deciding what kind of development

programmes they want, and thus the kind of transformations that will
take place, is not only a generally good thing. It should be understood as
necessary for genuine human development.
In the following section, I contrast the conventional development
and human development approaches to demonstrate the directly
transformative aspirations of the latter. Section II introduces the
concept of a personal identity capability. Section III analyses how
neglect of the transformational character of development can lead to
methodological and ethical problems. Section IV introduces the principle
of informed consent and its place in development practice. Section V
discusses the communitarian critique of transformational development.
Section VI analyses the necessary relationship between democracy and
human development.

I. DEVELOPMENT AS TRANSFORMATION
Sen has distinguished the capability approach to development from its
alternatives in terms of their respective focuses. He claims that
conventional approaches all focus, in one way or another, on providing
the means and circumstances for a better life (what I termed ‘capacity
building’). In practice this has meant the ‘fetishisation’ of indirectly
relevant features, such as economic growth, at the expense of what is of
direct relevance – the ability of individuals to live lives they have reason
to value. The capability approach addresses this tendency by subsuming
the logistical concerns of its alternatives within an account of
development that puts people at the centre (UNDP 1990, chap. 1).
Nevertheless, a focus on the means of development has one apparent
advantage. It allows an evasion of explicit consideration of important
ethical concerns about individual transformation. In contrast, the
capability approach must address these concerns directly.
Conventional ‘capacity building’ accounts of development policy
tend to represent, or implicitly assume a view of, the individual as
having an unchanging personal identity with respect to the development
process. This does not mean that development policies aren’t expected
to result in, or even intended to, change people’s choice behaviour. For
example, the recent ‘institutional turn’ in development economics has
brought attention to problematic social norms like corruption that
reduce and distort economic growth, while the human capital approach
emphasises how investing in children’s education and health can pay
114

off, both for them and for society as a whole. Development policies
directed at these goals (reducing corruption, increasing schooling) are
generally oriented to institutional reforms and incentives; that is, to
changing the constraints that individuals face (principally, budgetary
and informational), but not, directly, to transforming their values or
preferences.108 For instance, Kaushik Basu (as Chief Economic Adviser to
the Government of India) proposed making it legal to pay ‘harassment
bribes’, but not to receive them (Basu 2011);109 evidence from
randomised controlled trials has been used to suggest small
adjustments to the costs and benefits of schooling to make it more
attractive to parents, such as providing free school uniforms and free
lunches (Banerjee and Duflo 2011). Thus, in each case what is intended
is not the transformation of individual values, but the promotion of
certain behavioural patterns conducive to improving the functioning of
the economy and reducing material poverty.
A similar point can be made about the Basic Needs Approach to
development, which flourished briefly from the mid-1970s to early
1980s, in response to the perceived failings of GDP growth-based
approaches to development and income based views of poverty, before
being largely subsumed within the human development approach.
Although in theory the Basic Needs Approach was explicitly concerned
with democratic participation as well as with meeting minimum
requirements for goods and services like food, shelter, sanitation and
education, in practice democratic participation was often considered

108
“It is important to distinguish between genuine changes in values and those that
reflect alterations of relative weights because of parametric variations of the
determining variables” (Sen 2000b, 945; cf Becker 1996). For example, over the last 25
years, calorie consumption per capita in India has declined across all income groups,
and undernutrition and malnutrition levels remain very high, despite rising real
incomes (Deaton and Dreze 2009). Banerjee and Duflo note that when the prices of
cheap staples fall, or household budgets rise, as in India, the poor tend to buy less of
them, not more (Banerjee and Duflo 2011, 19–40). Instead they shift food budgets
towards more expensive tastier calories (such as more expensive grains, fats and
sugar) and increase spending on other ‘luxuries’ like festivals or radios. Banerjee and
Duflo interpret this as a shift in tastes allowed by increased budgets, rather than a
genuine transformation of preferences, and explain it by noting “the basic human need
for a pleasant life” and that poverty is very boring (37).
109
The logic of this proposal is that in the case of ‘harassment bribes’, where public
officials extract bribes for performing their mandated duties (such as taking crime
reports, issuing identity documents, and the like), the people who pay do so
unwillingly. If they were free to denounce an official, after successfully paying him to
do what he was supposed to have done anyway, they might well do so. Since all
officials would know this too it would significantly change their subjective expected
utility calculations around the action of demanding a bribe. As a result, this type of
corrupt behaviour should decline.
 115

separately if at all.110 Thus, in practice, the Basic Needs Approach was

mainly a technocratic enterprise concerned with the logistics of serving
externally identified universal needs. Hence also its general lack of
engagement with the people concerned in deliberating about what those
needs should be, which seems to have been perceived as patronising
and harmed its reception in poor countries (Stewart 2007, 15). The same
feature can be seen in contemporary examples of Basic Needs style
development policies, such as the Millennium Development Goals
project.
Thus, what I have termed ‘capacity building’ approaches to
development do not conceive the goal or processes of development to
be transformational. Instead they are focused on and justified by their
concern with increasing people’s capacity to live the lives they already
have (whether that is understood in terms of meeting their basic needs
or their personal consumption preferences111). This aspect is considered
an advantage by one prominent communitarian critic of development,
Stephen Marglin, who begins a book on the subject by stating,

Our criticism is directed against development and modernization,

not growth. The chapters that follow have nothing to say against
longer life-spans, healthier children, more and better-quality food
and clothing, sturdier and more ample shelter, better amenities. Nor
is any criticism levelled against the luxuries that people buy when
their incomes grow enough to permit discretionary purchases, such
as the radios and television sets that one sees even in very poor
Third World villages. (Marglin 1990, 1)

What Marglin seems to object to is not development conceived of as

capacity building, but development conceived of as transformation,
specifically that which, in his view, sets out to produce ‘modern’ people
in the Western model of individualistic rationality. The capability
approach’s transformative orientation, individualist ethical focus, and
concern with rationality (in the sense of ‘reason to value’ analysed in the

110
This point was noted in the first UN Human Development Report: “The basic needs
approach usually concentrates on the bundle of goods and services that deprived
population groups need: food, shelter, clothing, health care and water. It focuses on
the provision of these goods and services rather than on the issue of human choices”.
(UNDP 1990, 11) For other critical comparisons of the Basic Needs Approach with
respect to the capability approach, see for example (Sen 1984b, 513–515; Alkire 2005,
166–177; Stewart 2007).
111
This is the standard utility function - ‘Homo economicus’ - view in which an
individual is represented as having a given set of preferences that are specifically ‘their
own’. For a critique of the circularity of this account in terms of personal identity over
time, see John Davis (Davis 2003; Davis 2011, chap. 1).
116

previous chapter) therefore makes it a particular target of this strain of

communitarianism.112 I will return to this communitarian critique below
(section V). For now I will turn to showing that Marglin’s analysis of the
capability approach to development is largely correct: it is committed to
transforming people and not merely to helping them better live the lives
they now have.
But first it is necessary to question whether capacity building
development (mere ‘growth’) is actually neutral with respect to cultural
identity. As Denis Goulet, a pioneer of development ethics, notes,

The experience which villagers in traditional societies have of what

Westerners call technical or economic progress is that the values
which matter most to them – religious institutions, local practices,
and extended family solidarity networks – fall apart under the
impact of technology, the monetization of the economy, and the
specialization of tasks. (Goulet 1992, 468)

Thus, even development focused only on helping people to better

live the lives they now have does have transformational effects on those
concerned. It just does so indirectly and without evaluation, direction,
or accountability.
In contrast, the human development approach is committed not only
to making people’s lives go better, but to transforming them. For
example it sees education as important not merely for increasing
economic productivity, but also for directly transforming the lives of
individuals (their capability for practical reasoning, their social status,
the way that what they read may change them, and so on) and,
indirectly, for transforming society (by enhancing the inclusivity and
quality of public debate about social norms like fertility) (Sen 1997b).
Making this explicit directs our attention to how the capability approach
understands and represents individuals – in terms of agency, not
preference orderings or sets of needs.
Unlike conventional development the human development approach
deliberately sets out to transform people’s lives. The way the capability
approach represents individuals, in terms of their location in the
capability space, reflects this, since changing (improving) a person’s
location is the goal of development. Although the capability approach is

112
As another communitarian critic, Frédérique Apffel-Marglin puts it, “In all its
versions, the capability approach embraces individualistic modern rationality (in
the…Cartesian sense) as a universal and an indispensable tool for achieving increased
capabilities and thus increased well-being”. (Apffel-Marglin 2010, 210)
 117

often said to be concerned with ‘expanding’ individuals’ capability sets

(including by Sen (e.g. Sen 1989a)) the word ‘expansion’ may be
misleading, since the capability approach is actually concerned with
enhancing the quality of options people have access to rather than
merely adding to them.113 This is an important distinction, because
development doesn’t simply provide people with additional options on
top of what they already had, such as for industrial sector formal
employment as well as their present self-employed craftwork. It is quite
likely that the new options will permanently displace old ones. (For
example, the same factory that employs people in mass-producing
clothes is also likely to reduce demand for the more expensive,
traditionally made clothes, and hence the capability for self-employment
in that craft industry as it was originally configured.) This is an
important reason to ensure that the new options really are better than
the old ones, i.e. that one has development rather than merely change.
‘Capability enhancement’ may therefore be a better term for what the
capability approach is concerned to achieve.
Development understood as capability enhancement has
transformative implications for individuals’ lifestyles and values, and
for societies in general. Achievements of some functionings lead to a re-
evaluation and re-ranking of other functionings.114 For example,
increased individual mobility may change how an individual - and
society as a whole - values community and family life and associated
norms. Development will likely also change the definition of particular
capabilities, such as what constitutes adequate health-care or literacy
(often making their requirements more substantial). To give a more
concrete example, women’s literacy is strongly associated with increased
empowerment and substantial social effects on family life, including
lower fertility and a more equal intra-family distribution of resources
between males and females (Alkire 2005, 255–271; 294–6; Sen 1999a,

113
Though freedom of choice between good quality options is also important in Sen’s
account, and so having several good options from which to choose is better than only
one, even if they are all of the same quality.
114
The previous chapter discussed the problem that the adaptation of values, desires,
and aspirations to circumstances poses for the autonomy of those concerned. The
claim that transformative development can respect the autonomy of those concerned
thus requires that such transformations in people’s values are not merely reactive, but
consistent with or the direct conclusions of critical reflection (i.e. they are changes that
the impartial spectator could go along with). Offering reasonable assurance that this
will be so provides further justification for supporting and promoting the personal
identity capability, and guaranteeing the principles of Free Prior Informed Consent and
Democratic Development.
118

198–9). Altogether then, the capability approach to development expects

(and promotes) multiple changes in individuals’ and society’s self-
understanding, values, and ways of life; some intended, others
unintended but foreseeable, and yet others that are more or less
unknowable.
The value of a person’s capability set is understood in terms of her
effective access to functioning combinations she has reason to value.
Thus, an individual’s capability depends not only on her commodity
entitlements and ability to convert them into functionings (the logistical
aspect of the capability approach), but also on her reasoned valuation. 115
What kinds of life an individual has reason to value will depend on her
concerns and interests, which may be quite different from other
people’s, though still influenced by local social norms and
arrangements. The important point to note is that this valuation is
dynamic and endogenous to the development process itself. That is, as
part of the development process individuals are expected to change their
views both about which specific capabilities matter and about what
constitutes a good life.
How can people be understood as changing in this way while
remaining the same? The capability approach is evolutionary in that
having any particular set of capabilities opens the door to acquiring
additional capabilities. For example, when a person achieves a certain
level of education, they then have opportunities for employment they
previously did not have. Thus the person with basic literacy in her
native language who exercises her capability for further education and
chooses to pursue an advanced degree in, say, ancient Sanskrit
literature. Pursuing this option will shape her future capabilities for
work and leisure, while the commitment it requires may also reduce her
access to other, plausibly valuable, kinds of life and specific capabilities
that were originally open to her (cf Livet 2006). That is, individuals are
understood as autonomous agents who are engaged with development
in a co-evolutionary way. There is no claim here, as there is implicitly in
the capacity building approaches, that development is something that
happens apart from individuals and their values. Unlike those accounts,
the capability approach cannot avoid directly addressing the ethical

115
Of course, for many evaluative purposes specific lists may be used, for example
concerning a threshold for what is generally agreed to be severe poverty (Alkire and
Santos 2010a), or to focus on a particular issue like gender inequality (Robeyns 2003).
Nevertheless the foundational concern of the capability approach is with individuals’
capability to live the lives they have reason to value.
 119

challenges involved in combining autonomous agency and

transformative development. The first part of addressing that challenge
is to give a clearer account of agent autonomy. How should the
capability approach conceptualize the ability of people to be the author
of their own lives, to change while remaining the same person?

II. THE CAPABILITY FOR PERSONAL IDENTITY

In Development as Freedom, Sen contrasts the “agent-oriented”
capability approach with conventional approaches to development in
terms of treating those concerned as agents rather than merely as
patients. As he puts it, “With adequate social opportunities, individuals
can effectively shape their own destiny and help each other. They need
not be seen primarily as passive recipients of the benefits of cunning
development programs” (Sen 1999a, 11). Sen defines an agent “as
someone who acts and brings about change, and whose achievements
can be judged in terms of her own values and objectives” (Sen 1999a,
19). This is a good start, but some further conceptual elaboration is
required to relate the centrality of this idea of agency to the specific
issue of transformation.
A first step is to move from speaking of agency in general to
considering personal autonomy in particular. Autonomy is construed
here as self-government, of acting according to a self-chosen plan.
Autonomy thus refers to something more specific than the general
natural human capacity for agency described by Sen, of holding and
acting on one’s own values. Autonomy may be understood in terms of a
capability to plan one’s life according to one’s values. It involves
counterfactual analysis and intertemporal decision-making. For
example, a smoker may want a cigarette, and yet at the same time want
to be the kind of person who doesn’t want to smoke, and may make
plans and take actions to bring that other person about (Sen 2002e, 18).
Yet that other future self would not appear from nowhere but would be
linked to the present self by the autobiographical narrative that can be
told in either direction in terms of option paths chosen. Thus
understood, autonomy has complex requirements, including personal
faculties for practical reasoning (imagination, calculation, making self-
binding commitments); powers of intervening in the world; reliable
information; and a predictable environment. It therefore seems well
suited to capability analysis, i.e. in terms of the multi-faceted
120

relationship between persons and resources that determines effective

access to valuable functionings.
A second step is therefore to analyse personal autonomy in terms of
a personal identity capability.116 This move allows us to represent the
aspect of agency with which we are concerned within the capability
framework, and also to evaluate its adequacy and scope for
enhancement with all the nuance that comes with working within that
framework. First, as I have noted, the human development approach
represents individuals as collections of capabilities which change over
time. Introducing a specific personal identity capability to reflect on
what we have and who we wish to become, and to manage our
capability sets accordingly, allows us to make sense of how people can
change for the better or the worse over time as our capability sets
evolve, rather than merely be changed. Individuals purposefully develop
their capability sets in particular directions and so their development
can be assessed with respect to those plans.
Translating our concern for personal autonomy into the terminology
of capabilities also clarifies its centrality, which may be missed when
talking loosely of the general importance of agency for the capability
approach.117 It has the character of a ‘meta-capability’ in two senses. It
underpins the coherence of agents over time and is thus necessary for
human development, that is, for capability enhancement with respect to
agents’ own values and objectives. And it is defined over bundles of
capabilities since those are the features by which agents are represented
at any one time, just as regular capabilities are defined over
combinations of functionings with respect to choice.

116
This may seem an unorthodox use of the term ‘capability’, since it includes a kind of
freedom about freedom. I argue that this is justified to the extent that it is helpful in
the disaggregated evaluation of personal autonomy and how it might be enhanced.
Since the capability approach is not a theory but a flexible framework for considering
the evaluation of advantage, such ‘unorthodox’ uses are not only legitimate but even
commonplace. These include Sen’s own concern with the rather opaquely named
political freedoms, economic facilities, social opportunities, transparency guarantees,
and protective security in Development as Freedom (Sen 1999a, 10), and his
incorporation of freedom the definition of ‘refined functionings’ (like fasting) from
which one can choose; Nussbaum’s inclusion of practical reason (as an “architectonic”
capability) and control over one’s environment on her list of central capabilities (e.g.
Nussbaum 2011a, 33–4); and Erin Kelly’s recent argument for understanding “public
reason as a collective capability” (Kelly 2012).
117
A further advantage of this formulation is that the line of analysis it advances will
not be mistaken as a contribution to the extensive philosophical literature on
autonomy.
 121

Second, translating personal autonomy into a capability for personal

identity permits us to analyse it as we might any other capability. There
are relevant resources, such as access to credible truthful information,
time free from the demands of work and duty (leisure), and a stable
enough environment to permit long-term planning. There is the
heterogeneity of individuals, in terms of their internal and combined
capacities for practical reasoning, their social environment, powers of
intervening in the world, and so on. Just as with other capabilities, such
as the capability for appearing in public without shame, being capable of
managing one’s capability set over time is not a yes or no issue, but can
be met to one degree or another. Capability analysis can be employed in
the usual way: to identify inadequacies, diagnose their specific causes,
and recommend how they might be ameliorated. For example, to show
that those concerned have not understood what a development project
is about; to determine that this is because the information they were
given was in the wrong language; and to recommend action to ensure
they are properly informed.
Given its centrality it would seem that enhancing individuals’
capability for personal identity should be a development priority. Here
let me give just one example of its general relevance that also clarifies
its advance over the general concept of agency. In Poor Economics
Abhijit Banerjee and Esther Duflo discuss the consumption and saving
behaviour of the poor and the puzzle of why the poor don’t take up
opportunities to save for their future. They note that many of the poor
don’t feel a sense of control over their lives, that their choices can ever
add up to a different better future, and suggest, “Perhaps this idea that
there is a future is what makes the difference between the poor and the
middle-class” (Banerjee and Duflo 2011, 229).
But this not because the poor lack choices or rationality. In fact they
face a bewildering number of potentially life-altering choices every day,
which people in the rich world rarely have to trouble themselves with,
from navigating unreliable and untrustworthy health and education
systems to securing safe drinking water (Sen 1992, 62–4; Duflo 2012
Lecture 1).118 Merely surviving deprivation requires rational agency in the

118
"The richer you are, the less responsibility you need to take for the basic
constituents of your life (retirement savings, clean water, immunizations) because
everything is taken care of for you. While the poor have to be responsible for every
aspect of their lives, if the rich make no decisions and let the status quo obtain, they
are likely to be largely on the right track. For most of the poor, if they do nothing, they
are on the wrong track." (Duflo 2012 Lecture 1, introduction)
122

general sense. Rather, the character of their deprivation here concerns

the absence of valuable ‘options’: choices that they are able to relate to
achieving a better life.119 The people surveyed by Banerjee and Duflo
appeared unable to systematically link their choices over the
alternatives open to them to feasible pathways to the kind of life they
want to live. In other words, their capability to manage and enhance
their own capability set is often distinctly limited, and this is quite
compatible with agency in the general sense, and even with substantial
freedom to make life-altering decisions.
This section has clarified the capability approach’s concern with
agent autonomy and suggested how that may be analysed within the
capability framework. Yet individuals obviously face great difficulties in
making rational choices about their future. For example they may lack
adequate information about the outcomes of developmental choices,
which are nevertheless irreversible. That problem seems most severe
when their own values, and thus their frame of reference, are likely to
change over the course of development. In the following section I turn
therefore to discuss the issue of intertemporal choice by examining
what I call the ‘autonomy critique’ of development. Do the subjects of
transformational development give their informed consent to the
changes that occur as a result of or as a goal of development? Or are
these changes, in one sense or another, imposed by outsiders who know
best?

III. IS DEVELOPMENT GOOD FOR THE INDIVIDUALS CONCERNED? THE

AUTONOMY CRITIQUE
In a co-written paper, Sen and Nussbaum identify two ways in which
development can undermine values (Nussbaum and Sen 1989, 300). First
there is what they call “object failure”: when structural changes make
traditionally valued goods or ways of life more expensive or difficult to
obtain. For example, industrialization may increase the relative costs of
labour intensive goods such as religious rituals or traditional dress, or
land reform may make many of the traditional ways of life of nomadic

119
Personal identity capability deprivation is of course not only related to the absence
of valuable options. Adaptation, invidious social norms, stress, or mental illness may
also be causally implicated, and would remain significant outside the circumstances of
severe material poverty with which Banerjee and Duflo were concerned.
 123

peoples impossible. This may be experienced directly as a sense of

loss.120
Second, and analytically much more difficult, is what they call “value
rejection”: when people turn against their old values. For instance,
Buddhist Japan was once religiously vegetarian but meat-eating became
ubiquitous with Japan’s modernisation and rise to opulence; in Nepal
the nuclear family household model is rapidly displacing the extended
family; and so on. Such value transformations may also be a direct goal,
rather than merely a foreseeable byproduct, of development. In
Development as Freedom, for example, Sen argues for extending
education and opportunities for paid extra-household employment to
women in part because it transforms their lives, in terms of their
empowerment and emancipation and also social attitudes to family size,
typically resulting in a dramatic reduction in fertility (Sen 1999a, 198).
Value rejection is methodologically problematic because the frame
of reference shifts, so whether the change is a loss or a gain for those
concerned cannot be read off from either their antecedent or
subsequent values. It is ethically problematic, in terms of paternalism,
because it opens a space for a stable external authority to provide that
evaluation, an authority which, by definition, cannot be held accountable
to those concerned. When the same authority is concerned with
deliberately engineering that value rejection, all the elements are in
place for paternalistic development policy.
To identify more specifically how such paternalism may occur
inadvertently even in well-intended (beneficent) development projects I
analyse it in terms of the relationship between distinct intertemporal
selves of the individuals concerned. The ‘autonomy critique’ I identify
parallels a famous criticism of utilitarianism for disregarding the proper

120
Note that these goods can be undermined by development in two ways that bypass
the individual freedom of some while perhaps resulting from the free individual
choices of others (Marglin 1990). First there are externalities to development, such as
the factor price changes across an economy noted above, that may make traditional
goods more difficult or expensive to obtain for those who still value them. Second
there may be scale requirements to many traditional practices, like putting on a
festival or speaking a language: unless a certain number of other people, or perhaps a
certain proportion of a specific population, value those things and continue to take
part, they won’t be available anymore. For example the travelling storytellers who used
to go from village to village in India reciting and dramatising traditional tales like the
Ramayana (which takes several nights) need a paying audience of a certain size to
continue. If enough people switch to other sources of entertainment, such as
television, or to jobs which don’t allow them to stay up all night, then the remaining
audience is insufficient to provide the storytellers with a livelihood. I will come back to
this issue in discussing the ‘communitarian critique’ of transformational development.
124

boundaries between persons in its use of sum ranking (cf Rawls 1999,
24; Sen 1999a, 57). States of the world are assessed only in terms of the
total sum of welfare (however defined), and possible states are ranked in
terms of desirability from highest to lowest scores. In this approach
snapshots of social welfare are taken at different times and if aggregate
welfare at time2 is higher than at time1, then welfare is considered
improved, even though the welfare of some particular individuals may
have declined quite severely. With sum-ranking, the welfare of some
individuals may be sacrificed for aggregate improvement, and this is
part of what is generally recognized to be ethically problematic about
conventional economic development programs. When a dam is built in a
rural area to provide hydro-electric power for cities, it seems
questionable to call the results for those displaced from their homes,
communities, and livelihoods ‘development’ since their lives have been
made worse (cf Roy 1999, pt. II).
The autonomy critique raises similar questions about the sacrifice of
some individuals’ welfare for the sake of others, but focuses on
respecting the interests and values of inter-temporal selves within the
life of the same individual. That is, development is often understood
and evaluated as an end-state: the production of people with certain
features, whether that be greater opulence or an expanded capability
set. For example literacy or morbidity statistics are compared before
and after a policy intervention. The problem is that this comparative
statics approach neglects the dimension of ‘becoming’, including the
processes by which an outcome is brought about and whether these
respect the personal autonomy of those concerned.121 Extending the
evaluation of individual advantage to the capability space (i.e. to
incorporate non-pecuniary ‘beings and doings’ such as empowerment
and literacy) enriches such analysis but does not address this dimension
of becoming.
The ethical force of the autonomy critique is to highlight the
possibly illegitimate conflation of a person’s interests and values at
different points in time. It is motivated by a concern to justify and
assess development with proper regard to each person before as well as
after she takes part: no-one should be ‘forced to be free’, even for their
own future self’s sake. Even if it is generally agreed that the ‘developed

121
Note that incorporating the evaluation of process into consequentialist analysis has
long been a theme of Sen’s approach to evaluation in general (Sen 2000a) and
development in particular (Nussbaum and Sen 1989).
 125

life’ is better - and even if the ‘developed person’ herself endorses that
ex post (as in Mill’s famous Pushkin and pushpin example), there is a
troubling circularity in assessing and justifying development only or
mainly from the single perspective of the conclusion. Firstly one may
query the ethical justification for development if the ex ante evaluation
and concerns of that person are ignored or neglected. And secondly it is
hard to see how one could adequately evaluate the benefits or failings of
these changes to that person without considering the perspective she
started from as well as where she ended up.
A nice example of the problems this raises may be found in Sabina
Alkire’s pioneering work in Valuing Freedoms (Alkire 2005) on
operationalizing the capability approach by developing a capability
based approach to the cost-benefit evaluation of development projects.
Alkire considers various exemplary NGO projects in Pakistan, such as
rose cultivation and goat raising, and shows how the capability
approach allows a wider range of significant impacts to be included in
evaluation than merely financial returns.
However in one Oxfam project, teaching adult female literacy, inter-
temporal problems appear (Alkire 2005, 255–271; 294–6). Alkire relates
that the program was promoted to women and taken up by them (with
the permission of their fathers/husbands) principally on the basis of
claims that it would make them more employable. Oxfam’s other aim of
empowering the women was not emphasized or even necessarily
explained (though the choice of teaching methods inspired by Paulo
Freire suggests its centrality). There were, however, no job opportunities
for the graduates in the local area (because suitable workplaces would
not employ women). Nevertheless Alkire found that the project “had a
fundamental and transformative impact on the women students” (Alkire
2005, 256): they reported increased empowerment and greatly valued
this, despite it not being one of their original reasons for participating.
What seems problematic about the literacy project is not its
transformative goals, but its structure, which raises questions about
both legitimacy and assessment. One can question the legitimacy of
projects which recruit people by appealing to interests which will not be
fulfilled. Are those people being properly respected as bearers of ends,
or are they being used as means for the furtherance of the interests of
their future selves?122 Indeed, Alkire herself is somewhat troubled by the

One further argument for the structure of the project could be that if enough
122

women in the local area were to become literate, supply would create its own demand:
126

possible duplicity or paternalism in how the literacy program was

presented as opposed to how it worked. She notes that her capability-
based approach to the evaluation of development projects “does not
provide a way to distinguish activities which use informed consent from
activities in which consent is built during the process” (Alkire 2005,
296).
One can also be sceptical of cost-benefit analysis in such cases since
the valuational frame of reference is not constant and it is unclear which
set of values truly represents the individual concerned. From the
perspective of the women at the point they agreed to take part in the
literacy programme, it might be seen as a failure, or at least less of a
success, because the main projected benefit – jobs – did not appear,
while the costs in terms of time were significant. If those women had
known that there would be no jobs at the end of their education, they
might not have agreed to participate (and their male guardians might
not have permitted them to). Yet from the perspective of the women
after completing the programme, an unanticipated result –
‘empowerment’ – was perceived as a central and significant benefit.
These women might have been very glad that they had decided to
participate, even if things turned out differently than they had been
prepared for.
Putting the issue this way, in terms of present and future selves, may
seem obtuse, but it points directly to significant ethical issues. As Alkire
notes, focusing entirely on the ex post valuations of the women and
ignoring their ex ante valuations would allow all sorts of intensive
political or religious indoctrination programs to be justified in exactly
the same way as beneficent development programmes: the people who
go through them will claim to value what they have become.123 This
suggests that ex post evaluation is very susceptible to domination by
external policy-makers, substituting their own values for those of the
people subject to their policies and then attempting to bring such
people about.

social norms would shift and job opportunities would appear (Alkire 2005, 280). But
this still means using the present students as a means to an end in some sense, and in
any case the scale of the change in women’s literacy that would be required would
seem far beyond the capacities of that NGO project to achieve.
123
This can be generalized further. Neglecting how people’s value transformations
come about – for example, whether they are ‘brought about’ by others – would seem to
leave development programmes open to the same general critique of adaptive
preferences on which the capability approach is itself (partly) founded. In the
terminology of the discussion of the previous chapter, it is consistent with valuational
neglect, conflating an outcome (desiring) with a process (valuing).
 127

The autonomy critique questions the standard practice of assessing

development by comparing how well people are doing before and after
an intervention. Development work founded on the capability approach
improves on conventional evaluations by including non-pecuniary
aspects of how well individuals’ lives are going. Yet the autonomy
critique notes the methodological problem that a shifting valuational
framework poses to any cost-benefit evaluation. It relates this issue to a
second and more fundamental problem in determining the legitimacy of
the process of transformation. Only if individuals are transformed in a
way that respects their personal identity capability to manage their lives
over time, can their ex post perspective be reliable in informing us as to
whether their lives have been genuinely improved rather than merely
changed. A key tool for ensuring this, offering practical guidance as well
as conceptual resources, is the principle of Free Prior Informed Consent
(also discussed by Alkire 2005, 146–8; 150–1).

IV. THE PRINCIPLE OF FREE PRIOR INFORMED CONSENT

The problem with paternalism is that it ignores actual people’s forward-
looking evaluations or treats them strategically, as obstacles to be
overcome. That is inconsistent with the central ethical orientation of the
capability approach, in which development is to be considered not as
something that is done to people, but as something in which people are
involved in authoring for themselves. Genuinely human development
must engage with individuals as autonomous agents in managing their
own transformation. An important way of ensuring this is to
institutionalise the principle of free prior informed consent throughout
development practice.124 This principle provides a robust framework that
those concerned with planning and carrying out development
programmes can refer to in considering the practical requirements of
“agent-oriented” development in different kinds of cases. It can clarify
what follows from respecting agency, in terms of a requirement to
124
Although most systematically developed in the area of bioethics (with respect to
medical treatment and research involving human subjects) the application of the
informed consent principle is not limited to dealings between institutions and
individuals, but has also recently been extended and adapted to cases where large
numbers of people are involved, such as the reform of health systems (Daniels 2006);
economic policy making (DeMartino 2011; Blomfield 2012); and the rights of
indigenous peoples in development contexts (Goodland 2004). In such cases it is the
self-determination of a community, rather than the autonomy of an individual, that is
at stake and legitimate decision-making processes will vary from the bioethics case (for
example through majority voting decision-making). I consider such extensions below,
under ‘democratic development’.
128

respect, protect, and support the personal identity capability of those

involved to plan their lives in accordance with their own values. It also
provides safeguards against well meaning paternalism, by which the
values of development agents are, perhaps inadvertently, substituted for
those of the people they are trying to help.
According to the principle of ‘free prior informed consent’, the
people concerned should be adequately and truthfully informed in
advance of the overall aims of a project, its benefits and costs to them
(and the degree of uncertainty about these), and should be free to say
yes or no. In the bioethics literature, in which this principle has been
most thoroughly analysed and debated, it has been justified by a
number of distinct though overlapping moral concerns, including
protection of well-being; autonomy (the most prominent); prevention of
abusive conduct; trust; self-ownership; non-domination; and personal
integrity (Eyal 2011, sec. 2). Each of these justifications emphasises a
distinct moral concern about the issues and relationships involved. For
example, the justification in terms of trust points to the long term
requirements of persuading people to take part in the kind of
relationship which such treatment/research requires. The justification
in terms of preventing abusive conduct is concerned with how the
simple obstacle of gaining informed consent can support the
effectiveness of codes of conduct in governing the behaviour of doctors
and researchers.
Each of these moral concerns, or close analogues, is also of potential
relevance to the ethical conduct of development projects, though
different kinds of situation will likely call for emphasising different
moral aspects. For example, in many cases the beneficent intentions of
‘development agents’ may be more questionable than in the Oxfam
literacy project, and informed consent might then be particularly
important as a tool to prevent abusive conduct. Indeed, the informed
consent principle has thus far been most clearly formulated and
institutionalized in the development context with respect to protecting
indigenous peoples from exploitation in the name of development.125 Yet
genuine human development policy requires more than good intentions.
125
It is recognised that indigenous peoples are particularly vulnerable to coercive
exploitation of their lands and resources, even when that goes against national laws
and official policies, because of their degree of political and social exclusion in many
countries. The relative simplicity of informed consent requirements (for example as a
requirement by the World Bank for supporting projects) may protect such
marginalised groups from abusive conduct in a way that more complex, under-
enforced or politicised national laws may not (cf Goodland 2004).
 129

It also requires respecting, protecting and supporting the personal

autonomy of the individuals concerned. The wider institutionalisation of
the informed consent principle in development practice can play a role
in bringing this about. Understanding personal autonomy in capability
terms can help to clarify its requirements.
As I noted, focusing on personal autonomy rather than agency in
general brings out the importance of factors other than holding values
and choosing freely, such as being adequately informed. Free prior
consent is insufficient for meeting the requirements of personal
autonomy, because, while it guarantees a choice between alternatives,
without adequately informing people about those alternatives it fails to
provide them with options that they can relate to their ideas of a better
life. People may be provided with a selection of different levers to pull,
but yet be unable to systematically link their choice of levers to plans
for their future lives. Thus, what might be seen as the disrespect for
autonomy involved in inadequately informing the literacy project
candidates is that it undermined the women’s capability to govern their
own lives. They were given a choice of levers and tried to use the
opportunity to pursue the option of getting a paid job. But it turned out
that the lever wasn’t connected to that option after all.
We are concerned in capability analysis not only with the resources
that people have access to and their characteristics, but the relation in
which individuals stand to these (their ‘personal utilization function’).
Understanding personal autonomy in terms of a personal identity
capability allows the multi-dimensional character of its determinants to
be scrutinised. The principle of free prior informed consent can be
analysed in relation to that personal identity capability, in terms of
prescribing minimal threshold requirements for autonomy with respect
to the development project concerned. Where that capability to manage
one’s own development appears insufficient, the informed consent
principle requires us to analyse why this is so and what can be done
about it. We should ask not only whether the information people are
given is adequate (truthful, relevant, and including caveats), but whether
they have been adequately informed (understanding). We should ask not
only whether people have choices, but whether these constitute valuable
options that they can exercise to improve their lives.
Sometimes it might seem that informing people about the
consequences of their decisions is impossible and that the duty of the
development agent is rather to persuade them to make the right
130

decision. Even if those running a literacy programme know the likely

outcome – empowered women who value their transformation – it might
seem obvious that potential recruits would be incompetent to
comprehend or properly evaluate that outcome in advance (cf Alkire
2005, 295–6). But it is not so obvious to me. For example it is already
established practice, as part of the requirement to adequately inform,
for members of indigenous peoples to visit completed dams and mines
of a similar scale to those proposed for their area and talk to those
affected by them, in order to help them think through how similar
projects would change their own lives (Goodland 2004, 68). Likewise,
prospective students can meet graduates to see for themselves how
literacy does and doesn’t change lives. The significance of the informed
consent principle here is to require development agents to recognise
their positive moral responsibility to ensure that their clients are
adequately informed as well as adequately empowered to make an
autonomous decision about their development.
Supporting as well as respecting the autonomy of the subjects of
development requires astute judgement by development agents that
cannot be straightforwardly codified, for example with respect to the
control of husbands and fathers over enrolment in the women’s literacy
project. Embedding the principle of informed consent in the practice of
development is not meant to make this any easier. Rather, it takes the
form of an explicit ethical commitment to meet this challenge. Its
contribution is to bridge the gap between the theoretical understanding
of the requirements of genuine human development and the goals and
design of development practice by spurring and directing ethical self-
scrutiny.
My analysis so far has focused on the relationship between
individuals and development programmes, and the procedural
requirements that follow from that for legitimate human development.
Yet it is of course the case that most transformational development
programmes take place at the social not the individual scale. The issue
of autonomy arises again in two senses. Does social transformation
undermine individuals’ self-identity, as constituted by their social roles,
affiliations and ways of life? And, how can legitimate decisions about
social transformation be made by a community? In the remainder of this
chapter I therefore turn to considering the requirements for respecting
autonomy in development at this more complex group level, starting
 131

with an examination of the ‘communitarian critique’ of development as

transformation.

V. DOES DEVELOPMENT DESTROY PEOPLE? THE COMMUNITARIAN

CRITIQUE
Development brings multiple changes in individuals’ and society’s self-
understanding, values, and ways of life, some direct and more or less
foreseeable, others indirect and complex. From a communitarian
perspective such radical transformations of ways of life and social
values may be seen as an existential threat to the self-identity of those
affected. As the communitarian critic of development, Stephen Marglin
puts it, quoting from the New Testament (Mark 8:36; Mathew 16:26),
“What shall it profit a man, if he shall gain the whole world, and lose his
own soul?” (Marglin 2010, 288).
Communitarians claim that people’s self-identities are constituted
by their values, and that those values are ‘irreducibly social’, in that, like
a language they are constituted by the actions of individuals (speech
acts) but not reducible to them (Taylor 1995). Communitarians can agree
that individuals are fundamental units of moral concern, but the
irreducibly social nature of self-identity means that social structures are
more fundamental and thus have intrinsic value as well (Taylor 1995,
137). As one communitarian capability scholar, Sévérine Deneulin, puts
it,

Community is pre-existent to individuals. It is what gives meaning to

the life of its members and gives them identity, in the sense that it is
only from their attachment to communities that human beings draw
their moral development, their identity, and the meaning of their
life. (Deneulin 2008, 120)

From this communitarian perspective, development threatens

identity from two directions. By displacing and superseding traditional
“structures of living together”126 it removes the social scaffolding from
people’s self-identities. By directly transforming people’s values it
supercedes and displaces their core telos. Although communitarian
philosophers have different positive accounts of self-identity (such as
Alasdair MacIntyre’s and Charles Taylor’s, which I discuss briefly below),

126
Following Deneulin’s terminology, slightly adapted from Paul Ricoeur, in which this
is defined as “structures which belong to a particular historical community, which
provide the conditions for individual lives to flourish, and which are irreducible to
interpersonal relations and yet bound up with these” (Deneulin 2008, 111).
132

their general concern can be understood as forceful scepticism of the

individualistic free choice view of the liberal political philosophy
tradition (Bell 2012). Though I will not be going along with their positive
accounts, which I believe are unconvincing and excessively conservative,
I do think a transformative account of development can benefit from the
ethical scrutiny they bring.
Communitarians are strongly associated with the narrative account
of self-identity which requires a certain kind of coherence across an
individual’s whole life: a person is a person because they can tell an
intelligible story about their history and future in which they are
embedded. A person’s life evolves in the sense of an unfolding story, in
which the purpose becomes apparent to a reflective protagonist.
Narrative identity is understood differently by different writers (such as
MacIntyre and Taylor), but the danger is generally understood in terms
of a breakdown of the telos of a person’s life.
MacIntyre’s account is strongly historical and social and focused on
intelligibility. We are in effect given a life – a narrative determined by
historically contingent social roles, meanings and values – and our task
is to live it well, or to enact it (MacIntyre 1981, 211–2). Narratives are
thus teleological in that they present an intelligible direction and
structure for a life. And these narratives are social – individuals are only
ever co-authors – because individual stories are embedded in the ways
of life of a community and the goods they understand as valuable, and
those goods can only be understood and produced within those terms.
On MacIntyre’s account, then, development that changes the
community disrupts the coherence of an individual’s narrative by
changing the story - the end and the beginning can no longer be
understood as part of a coherent life narrative – and by changing the
context (of the structures of meaning of a community life) so that an
individual’s life is no longer intelligible to her. From the outside, one
could tell a story about how an individual came to live differently, but
from inside, the individual would no longer be able to tell an intelligible
coherent story of herself. The result: events keep happening in the
individual’s living, and keep causing new events – we still have
psychological continuity - but these are now just a heap of events,
ordered in time but no longer experiencable as a coherent self-
understood narrative.
While Taylor is also concerned with the importance of social
structures for an individual’s ability to lead a valuable life (Taylor 1995),
 133

as far as the narrative account goes he focuses on a person’s overall

thematic unity or telos over a whole life, and the contrast between self-
evaluation and an individualistic concept of free choice. Taylor’s
persons are more capable than MacIntyre’s of dealing with conflicts and
transitions so long as they don’t threaten their core telos. Taylor
contrasts his notion of ‘strong evaluation’ – of reflecting on what one
truly wants and values and considering which habits and concerns are
an essential part of one’s authentic self – with the emphasis on free
choice for what one most desires which he associates with modernity,
including liberal political philosophy and modern economics. On this
view, freedom refers to the ability to master one’s internal constraints
by critically evaluating one’s beliefs and values in terms of how well
they relate to one’s overall evaluative principles or telos. Choice is not
the relevant term here, but rather, with introspection, one comes to
realize that certain desires are not truly authentic and should therefore
be dropped. But the fundamental telos itself cannot be changed, only
better understood.
The capability approach to development places freedom at its heart
– the goal is to provide people with more valuable functioning
combinations from which they can choose. These are in effect more lives
from which to choose and construct for oneself. So there are two threats
from Taylor’s perspective. First, that the changes themselves will be too
radical and one will be presented with ways of life for which one’s core
orientation cannot cope. Second, that the kind of pick and choose
freedom development is concerned to produce is incompatible with
developing any thematic unity across a life.
It may not seem that communitarian political philosophers can have
much to contribute to the discussion of development ethics, insofar as a
strong interpretation of their claims suggests that any understanding of
development as transformation is misconceived and dangerous. For
example, individual freedom is constrained conceptually. Because
people don’t choose their values, they also cannot choose to change
them. Thus any transformation in people’s values is understood as
simply imposed on them from outside. Individual freedom is also
constrained structurally. Because social goods depend on the actions of
individuals to sustain them, an individual who did (try to) choose to be
different by exiting her traditional role would be putting the very lives
134

of others at risk since everyone’s self-identity depends on maintaining

the shared social structure.127
The valorisation of the status quo “structures of living together” by
communitarians has a strongly conservative orientation that would
seem to prohibit transformational development altogether.128 Such
valorisation is inconsistent with the rather plausible idea that the
flourishing of individuals can be assessed in ways that go beyond their
place in a social structure, and that such analysis reveals that many
people are living unnecessarily deprived lives. For such reasons it does
not seem controversial to reject strong versions of communitarianism in
thinking about development.
Nevertheless, understood as a critique of certain lacunas in liberal
political philosophy and political economy, communitarianism can
provide a useful service. First, it is successful in highlighting the social
character of development, in terms of the interdependence of individual
choices and the group character of many significant changes. This point
and its significance has long been accepted by Sen himself,129 though it is

127
For example, the communitarian political philosopher Michael Sandel frequently
refers to how non-standard choices and practices by individuals can corrupt the nature
of social goods and relationships like friendship or gift giving. Thus the freedom of
individuals to dissent from traditional norms poses a threat to the well-being of
everyone else in society (Sandel 2012). This is the same mechanism appealed to by
contemporary conservatives who argue that legalising ‘gay marriage’ would corrupt
and degrade traditional marriage.
128
Though attempts at communitarian ‘growth’ have been made, for example in
Bhutan’s famous ‘Gross National Happiness’ approach to development (since 1972).
The government of Bhutan has now developed evaluative criteria for this grounded in
‘Bhutanese’ culture and history, which differ from more orthodox understandings of
happiness firstly by including other dimensions than subjective well-being and
secondly by including non-individualist aspects such as harmony with nature and
concern for others (Ura, Alkire, and Zangmo 2012). Of course, the other thing Bhutan is
famous for is stripping civil rights and citizenship from up to 35% of its population
(the Lhotshampa), who do not belong to the dominant ethnic group, and the physical
expulsion of some 100,000 of those to refugee camps in Nepal and India (from the late
1980s). It is hard not to see a connection between Bhutan’s mistreatment of those who
are perceived as insufficiently Bhutanese and its promotion of an explicitly
communitarian national development programme. Nor does it seem co-incidental that
this path was embarked on before the introduction of democracy.
129
Thus, “Human beings live and interact in societies, and are, in fact, societal
creatures..... No individual can think, choose, or act without being influenced in one
way or another by the nature and working of the society around him or her (Sen 2002f,
79, 80).” And, "In judging development in the context of a culture, the values that are
supported and are sustainable in that culture provide an essential point of reference"
(Nussbaum and Sen 1989, 299–300).
 135

fair to say that it is not always reflected in the capability literature

which tends to represent development quite individualistically.130
In addition the communitarian critique can be seen as raising self-
identity over time as a central ethical challenge for development theory
and practice, and in casting that challenge in terms of self-
determination. People are socially situated and they do take their values,
including their core values from their context. But that does not mean
that people have no choices to make about their values, and thus their
identity. As Daniel Bell notes,

liberalism founded on the value of self-determination requires only

that we be able to critically evaluate our ends if need be, hence that
‘no end or goal is exempt from possible re-examination’ (Bell 2012,
sec. 2 original emphases; See also Sen 1999d)

Thus, the fact that individuals must always evaluate their situation
and their choices from some particular position or positions (rather
than some abstract ‘view from nowhere’) does not mean that they
cannot scrutinise those socially transmitted values and perspectives and
reject or re-weight them in light of that scrutiny and the knowledge of
other available perspectives.131 This capacity to consider one’s values for
oneself becomes a right if one takes seriously the claim of the capability
approach that development is concerned with the life that people have
reason to value.
Cultural traditions and values clearly can be reasonably valued by
people – they can withstand this scrutiny. Whether or not people are
poor in certain ‘basic’ capabilities (such as for being well-nourished or
participating in national politics) and come to see themselves as such,
they may be rich in others, such as participation in community life or

130
As De Herdt and Deneulin note, such an individualistic methodology “renders more
complex, but does not fundamentally challenge, the economists' view that people
consume a combination of goods subject to a budget constraint” (Herdt and Deneulin
2007, 180). Though the capability approach is not limited to such an individualistic
methodology (see for example James Foster and Christopher Handy’s work on ‘external
capabilities’ (e.g. Foster and Handy 2009)).
131
This capacity for self-reasoning was the focus of the previous chapter. Sen often
uses the example of Gandhi to illustrate its scope: “when Mohandas Gandhi decided,
after considerable reflection, to give priority to his identification with Indians seeking
independence from British rule over his identity as a trained barrister pursuing English
legal justice, there can be no question that he was consciously and firmly making a
choice.” (Sen 2004c, 17) A recent biography of Gandhi makes clear the sustained,
extensive and deliberate character of this self-transformation, from the prosperous
British trained, and tailored, lawyer who landed in Durban in 1893, to the ascetic in a
loin cloth who returned to India in 1914 (Lelyveld 2011).
136

cultural achievements. There may be some traditional values which are

either discovered to be unjust or are instrumentally incompatible with
achieving other capabilities considered more valuable. But that decision,
Sen has argued, is not determined in advance either by the authority of
tradition or technocratic calculations, but is rather dependent on an
exercise of “critical valuation” for the people concerned to decide
together through an open participatory process (Sen 1999a, 31–3; cf
Nussbaum and Sen 1989).
In the developed world vigorous public debates around the
significance and role of traditional values in our modern life are routine
(for example around the issue of ‘gay marriage’), and can often result in
public policies to preserve certain generally valued social goods even at
the cost of public funds and economic inefficiency (such as schooling in
minority languages or religious holidays). Likewise ‘social goods’ such as
community relations (and connected ‘external capabilities’ (cf Foster and
Handy 2009)) are considered important in development projects in the
developed world.132 Though the choices concerning social transformation
in under-developed societies may well be both starker and more urgent,
there is no principled reason why the people concerned should not also
take such decisions for themselves. One should, in general, be sceptical
of accepting any implication that deprived people lack the capacity to
make decisions for themselves and therefore cede the right to do so.133
Likewise there are likely to be other ‘social goods’, like the tradition
of travelling story tellers in rural India, which may be casualties of
development processes even though many people would like to keep
them. Here again an individualist use of the capability approach would
go astray. If one understands cultural liberty only in terms of the
aggregation of individual private choices, then the fact that fewer people

132
For example, the destruction of communities in urban development projects
intended to provide the poor with better housing has long been acknowledged as a
significant mistake and its lessons integrated in contemporary policymaking.
133
The 19th century promoter of liberalism as individual autonomy, J S Mill, for
example, proposed such a paternalist account of development. He had an unfortunate
conception of different national characters in terms of their position in the hierarchy
of civilisation. The status of backward peoples in this hierarchy was as of children:
incapable of self-governance. Thus Mill conceived of development as concerned with
producing a more civilised national character through paternalistic social engineering
(cf Pitts 2005, chap. 5). As he put it in the introduction to On Liberty (Mill 1859),
“Despotism is a legitimate mode of government in dealing with barbarians, provided
the end be their improvement, and the means justified by actually effecting that end.”
That contrasts strikingly with Mill’s faith in the capacities of women, once they were
freed from their conditions of subjugation, as discussed in The Subjection of Women
(Mill 1869).
 137

come to participate in such activities can be read off as a ‘social choice’

to end that tradition. But this would be to mistake a byproduct of
development for development itself, and exaggerate the capacity of
information about effective demand to inform us of social choices.134
Here again, Sen has argued, the people concerned should be able to
decide whether and how to support such goods in a fuller exercise of
social choice involving open participatory decision-making (Nussbaum
and Sen 1989; Sen 2004c, 20–1). If, through these democratic processes
of public deliberation, people come to value the capability to retain
access to these traditional social goods then that valuation should be
understood as determining what counts as development in this case.135
This section has argued that the communitarian perspective can play
an important role in thinking of development as transformation. While
its substantive ‘metaphysical’ account of the nature of personal identity
has little to offer directly, it raises ethical concerns which warrant
procedural guarantees about how social development may proceed.
Respecting individual agency in the sense of autonomy requires that
proposals for improving people’s lives take account of what the people
concerned actually have reason to value. In many cases this will require
participatory exercises of social choice so that communities decide for
themselves which trade-offs to pursue. In the following section I develop
this general case for the centrality of democratic deliberation by those
concerned into a side-constraint on development.

VI. THE PRINCIPLE OF DEMOCRATIC DEVELOPMENT

The human development approach takes a specifically democratic
perspective on development theory and practice that rests on a central
distinction between the constituents and determinants of the life one has
reason to value (e.g. Sen 1999a; Sen 2009a).136 The constituents of
development are the beings and doings that people have reason to
value, and are identified through an exercise in social choice, ideally
characterized by public reasoning and deliberation, by all those

134
There is a parallel here with Sen’s argument against using methods based on market
analogies to determine the social valuation of environmental goods (Sen 1995a).
135
This does not mean that individual liberty to hold different values from the ‘general
will’ established by participatory public reasoning is undermined, anymore than being
taxed to support public museums one doesn’t visit constitutes such an infringement.
136
This distinction was, I believe, coined by Partha Dasgupta, and employed in the
course of an argument that the considerations of the constituents of development by
Amartya Sen and others was superfluous (e.g. Dasgupta 2009). Although I disagree
with Dasgupta’s conclusion, the distinction is a helpful one.
138

concerned within a given society. The determinants of development are

those things (such as capital, commodities, institutions, laws, public
services, and so on) that are necessary for bringing about the kind of life
people have reason to value. They are provided by those agencies (such
as governments and development NGOs) which possess technical
knowledge of and command over such causal factors of development.
These development agencies may provide advice, for instance in the
form of critical external perspectives on gender justice, or factual
information about the determinants of development that may be
relevant to the assessment of feasibility. But their evaluations do not
determine what the constituents of development are, i.e. what people in
this society have reason to value.
This division of labour between the valuation exercise and its
technocratic implementation - values and facts - contrasts with the
standard development paradigm in which responsibility for valuation,
implementation and evaluation are combined in a single distinct group
of experts who are fully responsible for the conception and design of
development policy (typically central government in collaboration with
international development agencies). The democratic perspective taken
by the human development approach puts the policy exercise in service
to the valuation exercise so that social deliberation over the constituents
of development substitutes for their technocratic determination. As Sen
puts it,

The people have to be seen, in this perspective, as being actively

involved—given the opportunity—in shaping their own destiny, and
not just as passive recipients of the fruits of cunning development
programs. The state and the society have extensive roles in
strengthening and safeguarding human capabilities. This is a
supporting role, rather than one of ready-made delivery. (Sen 1999a,
53)

Of course, individuals will not always get their own way about their
society’s developmental priorities – the final outcome of the social
choice exercise may well need to be decided by majoritarian voting.
Nonetheless, the involvement of individuals in a forward-looking
democratic process of social deliberation over social goals is
qualitatively different from cases where either policy-makers decide
those goals directly or it is decided by one-off voting (mere choice). This
is because such exercises directly engage people as autonomous agents
 139

in the collective self-determination of their values and concerns. Such

exercises work through, not over, the personal identity capability of
individuals since they presuppose that agents reflect upon the
consequences of the options open to them, for themselves as well as for
society, in making their contributions to the participatory process. Thus
development pursued in this democratic manner is not paternalistic.
Rather, by making the individuals themselves the authors of change in
their lives, rather than groups of external experts, democratic
development is a necessary component of transformation without
paternalism.
This analysis of the demands of a truly “agent-oriented”
development thus provides a powerful supplementary argument for the
constitutive importance of democratic deliberation in human
development policy. Of course it is well recognized among proponents
of the capability approach that democracy and political freedoms are
intrinsically valuable dimensions of development in general, but this
recognition is not always extended to development policies in particular.
For example, local participation is often promoted in terms of
respecting agency in general as a goal, while I have shown that
participation should also be seen as respecting personal autonomy in
particular, as a side-constraint that is required for these projects to be
understood as human development rather than as something else.137
In addition to being required by this account of human development,
public reasoning and deliberation also have a positive instrumental
relationship with the ‘personal identity capability’ I identified, since they
provide important opportunities for its exercise and development. They
give people real options to choose between, and the ‘leisure’
(information, space, freedom, and other minds) to consider them
properly. That experience is likely to enhance individuals’ personal
identity capability in their private lives to reflect upon and deliberate
about who they want to become and how to get there (thus countering
the problem of adaptation discussed in the previous chapter). To the
extent that individuals reflect upon and change their own values or
identity through participation in public reasoning - for example coming
to see social norms concerning women’s fertility as invidious (Sen
1995b, 17) - they exercise their personal identity capability for

137
This analysis thus provides a fuller justification for Sen’s contention that “the need
for popular participation is not just sanctimonious rubbish. Indeed, the idea of
development cannot be dissociated from it”. (Sen 1999a, 247)
140

governing their own values and identity. Sen has argued for the
constructive value of public reasoning and deliberation. From the
perspective of the individual they might also be seen as
transformational.
This discussion of democratic development may seem both too
abstract and too idealistic to be relevant to actual development
programs and concerns. So it is worth showing local democratic
development can work in practice with a brief case study of a
development organisation that integrates it into its work.
Gram Vikas (“village development”) is a regional NGO active in
Orissa, India, which has pioneered a particularly ambitious approach to
the provision of water and sanitation in rural villages with high
populations of poor and low-status residents.138 Gram Vikas is
committed to a certain vision of development – “An equitable and
sustainable society where people live in peace with dignity” - which
clashes with commonly held local values in its concern for equity across
gender, status (caste) and income distinctions (Keirns 2008, 27–8).
Gram Vikas began its water and sanitation program in 1992 after
asking villagers what they needed most. Their respondents said they
needed a hospital because people were always sick. However Gram Vikas
noted that most of the diseases were waterborne, and their analysis of
the determinants of what the residents wanted – better health –
suggested an integrated program of drinking water and sanitation
provision would be much more effective. The program Gram Vikas
developed, and which has now been successfully taken up by more than
300 villages (covering over 150,000 people), has several interesting
features. Gram Vikas requires complete unanimity from all male and
female heads of households in a village before beginning; it insists that
every household receive the same quality facilities; and it requires each
village to set up committees to manage the water project and
maintenance.
Gram Vikas often takes more than a year to move from mooting a
project to beginning it because of its requirement for universal
commitment. In particular, from the start they insist that women, who
suffer most from the lack of private safe sanitation and piped drinking
water, must be part of a village general assembly that decides
(unanimously or not at all) to go ahead with the project and deliberates
about the details of implementation (such as equitable sharing of capital

138
This account is based on The Gram Vikas Experience by Pamela Keirns (2008).
 141

and labour costs and arrangements for operational management). It

takes some time for the general assembly to function properly,
particularly for the women to feel able to join in discussions. During
this time Gram Vikas facilitators, generally locally recruited, are
constantly available to answer questions about technical issues and the
NGO’s role. If the general assembly decides positively, the plan is
formalized as a contract between the village and Gram Vikas, with every
assembly member signing it.
The structure of this village-wide deliberative process conforms to
Gram Vikas’ commitments to income, status, and gender equity, since
all must be included in the deliberative process and all have a veto.
Gram Vikas justifies this not only in terms of its core values, but also in
practical terms that villagers often find more convincing. Ending
waterborne health problems has features of a collective action problem
in that unless everyone stops the practice of open defecation the water
supply will remain polluted. Furthermore, in order for the project to be
sustainable the village as a whole has to make a comprehensive
commitment to managing it (since Gram Vikas will gradually withdraw
over five years). The structure of this public deliberation not only
reflects Gram Vikas’ values, but helps to advance them because women,
the poor, and low status individuals are endowed with formal equality
for the first time, and in practice become more confident about
articulating their views and interests in public. Gram Vikas is as proud
of its achievements in advancing the dignity and standing of the
marginalized and building more harmonious communities – of
transforming values - as it is of the success of the water and sanitation
projects themselves.
Gram Vikas’ success is built on a respect for democracy, community
and autonomy that not only respects the personal identity capability of
individuals but enhances it. First, Gram Vikas respects the right of those
involved to collectively decide on their priorities, while offering
arguments about the benefits of the project for the village to consider,
and technical and financial support if it decides to go ahead. In this way
it seems to respect the division of labour between the constituents and
determinants of the life one has reason to value. It is not paternalist.
Second, it respects the requirements discussed in section V, that
proposals for improving people’s lives take account of what the people
concerned actually have reason to value. Gram Vikas engages in a
lengthy exercise in persuasion, not dictation. It uses a variety of
142

techniques, including organizing visits to villages which have completed

the project; culturally specific arguments to men that female family
members having to defecate in public is shameful; health arguments in
the form of public demonstrations by respected doctors; appeals to
village pride in having running water and private bathing rooms which
even richer villages don’t have; and so on.
In addition, communitarians may note, although Gram Vikas is
committed to transforming traditional social roles (and thus
communitarian personal identities), it is also committed to
strengthening community cohesion by getting villagers separated by
those traditional roles working and deliberating together on shared
village projects. The village assembly continues to meet once a month,
and elects a Village Executive Council (with proportionate representation
from men, women, and different castes) to manage the project. This is
formally registered as a village society so that it can manage funds and
also deal directly with the formal government development bureaucracy
and lobby for funding and manage other projects such as road and
school improvements. In addition many villagers become members of
sub-committees set up to address such issues as hygiene policing, water
system maintenance, education, and community resource development
(such as fish-farming) to pay for operation and management costs.
Third, it respects the personal identity capability of the individuals
concerned by requiring free prior informed consent. So although Gram
Vikas is committed to certain non-traditional values, at no point does it
attempt to substitute these for those of the people whose lives it is
trying to improve. First, it is transparent about its values, and that they
are a non-negotiable part of the program it is offering. Second, villagers
are able to reject Gram Vikas’ offer and for example pursue other
sanitation projects offered by the government. Third, it insists that
every head of household be included in the decision-making and be able
to articulate their concerns and objections. Fourth, and most
significantly, the problematic information gap – that forward-looking
agents might be simply unable to imagine and assess transformational
projects within their current evaluative framework – is substantially
bypassed since there now exist a great many villages with completed
Gram Vikas projects that can be visited.139

Of course this was not the case when Gram Vikas first began these projects. Then, it
139

depended on the close trust it had built with a handful of villages in its earlier
development work.
 143

I have argued that the paternalist implications of transformative

development can be addressed by institutionalising a respect for self-
determination. Individuals have the right to options and not merely
choices. Communities have the right to collectively determine their own
development paths. In this section these two aspects of autonomy,
individual and collective, were brought together in the framework of
‘democratic development’. Here, it is the people concerned, not select
groups of experts, who determine development goals, converting
collective understandings about the constituents of the kind of lives
people have reason to value into practical decisions about which
development paths to take.

VII. CONCLUSION
This chapter aimed to elucidate and address the ethical concerns
underlying the idea of development as transformation. I began by noting
that while conventional ‘capacity building’ approaches to development
evade these concerns, the human development approach cannot. Human
development not only aims directly at the transformation of people’s
lives, but it also claims to be an “agent-oriented” view. I believe this
implies specific goals and constraints for the practice of development
which have not so far been explicitly recognised.
In working these out, I noted that the human development
perspective assumes that an individual’s personal identity evolves with
development. I introduced the concept of a personal identity capability
to represent the understanding of personal autonomy this implied: the
ability to change one’s life, including one’s ideas about the kind of life
one has reason to value, and yet remain the same person. In doing so I
clarified the requirements of taking an “agent-oriented” view in the
context of value transformation. One can only evaluate whether people
are better or worse off, rather than merely changed, if they themselves
provide evaluative continuity in the form of an auto-biographical
account relating the paths chosen and their reasons for them.
I then showed that this somewhat abstract analysis of personal
identity has important implications for development practice. Analysis
in terms of personal identity capability can be helpful in identifying and
diagnosing ethical problems in the practice of human development,
which standard comparative static methods such as before and after
capability-set evaluations would miss. In consequence, I argued that
144

human development policy should incorporate procedural principles

that protect and promote the capability of those concerned to be the
authors of their own development: the principle of free prior informed
consent and democratic development.
The capability approach is generally understood as being founded
on respect for individual freedom and agency. What I believe to be less
well recognized is that it also implicitly relies on a conception of people
as evolving and able to reflect upon their personal identities and
individual development. This chapter has sought to make explicit the
theoretical and practical implications of this conception of the person.
 145

Chapter 6. Which Capabilities Matter for Social Justice?

Democratic Politics Versus Philosophy.

Amartya Sen has argued extensively against the sufficiency of resourcist

and utilitarian accounts for “judging individual advantage and in
evaluating social achievements and failures” (Sen 1999a, 285). And he
has argued positively that such judgements should consider individuals’
substantive freedom to lead the kind of lives they have reason to value
(their capability). Yet if we want to use the concept of capability to talk
about justice we must move from subjective to ‘objective’ criteria, we
must not only accept that the capability space is significant for the
evaluation of advantage, but also provide an account of how to decide
which capabilities have moral significance with respect to questions of
justice.140 Sen has argued that decisions about which capabilities a
society should value and guarantee are political, and should be made by
the people concerned, rather than by technocrats or philosophers (or
autocrats or religious leaders) (e.g. Sen 1999a, 81; Sen 2004a). And he
notes that such collective – political – decisions are of great practical
importance because of the significance of social arrangements for the
substantive freedoms of individuals (hence, Sen’s characterisation of
“individual freedom as a social commitment” (Sen 1999a, chap. 12; Sen
1990c)).
Sen clearly endorses a democratic approach to questions of social
justice in general and of which capabilities matter in particular. He has
argued for Democracy as a Universal Value (Sen 1999c) and makes many
references to John Rawls’ idea of ‘public reasoning’ throughout his
work. In Development as Freedom he argues that the capability for
political participation is “central”, “indispensible”, and “preeminent” for
development (Sen 1999a, 110, 287, 147–8). In a discussion about

140
Such criteria are ‘objective’ in the sense that they can be publicly justified. As T.M.
Scanlon argues in Preference and Urgency, “insofar as we are concerned with moral
claims that some interests should be favored at the expense of others in the design of
distributive institutions or in the allocation of other rights and prerogatives, it is an
objective evaluation of the importance of these interests, and not merely the strength
of the subjective preferences they represent, that is relevant” (Scanlon 1975, 658). Sen
would seem to agree: “In so far as some agreement is needed for the social framework
of human rights, the agreement that would be sought is not only whether some
particular freedom of a particular person has any ethical importance at all, but also
whether the relevance of that freedom meets the threshold condition of having
sufficient social importance to be included as a part of the human rights of that
person, and correspondingly to generate obligations for others to see how they can
help the person to realize those freedoms” (Sen 2009a, 367).
146

capability lists, Sen argued forcefully that their legitimacy and epistemic
relevance requires public deliberation (Sen 2004a, 78). Yet for all that, it
is fair to say that Sen has not been very specific about how this
democratic process should work.
Various authors have identified what may be called a ‘political gap’
in Sen’s account, though there seem to be two distinct diagnoses of
what that gap is. On the one hand Sen has been criticised for a lack of
idealism, in the sense of failing to provide a clear normative account of
how democratic decision-making should proceed (e.g. S. Srinivasan
2007).141 On the other hand he has also been criticised for excessive
idealism, in the sense of being naive about the Hobbesian nature of real
world politics, ‘democratic’ or otherwise (e.g. Shapiro 2011, 1259–62;
Stewart and Deneulin 2002, 63–4).142
In this chapter I will argue for interpreting Sen’s account of
democracy and public reasoning in social choice terms, as he himself
consistently argues. I will show that, seen in this way, Sen is both more
realistic and more normative about democratic politics than his critics
recognise. In line with his ‘comparative’ approach to justice, Sen
attempts, to navigate the tensions between the mutually antagonistic
perspectives of politics as struggle over interests or as applied morality.
Notably, Sen argues for the ‘politicisation’ of social injustices, from
hunger to missing women. That is quite different from the ‘moralisation
of politics’ project which Bernard Williams identifies as the central
approach of mainstream (Anglo-American) political philosophy
(Williams 2007). Sen recommends using politics to address moral
concerns. Critics of his ‘political gap’, such as David Crocker and Rutger
Claassen, argue for using moral theory to address moral concerns.
The following section outlines the ‘political gap’ in Sen’s writing: the
need for the political determination of which capabilities matter for
social justice and Sen’s failure to provide a clear account of how that
political process should proceed. Section II outlines Bernard Williams’
account of political moralism and its relevance here. Section III reviews

141
Thus, “Sen’s ‘silence’ on the substantive content of an account of justice is due in
large measure to his stringent emphasis on plurality, agency and choice; he turns to
democratic processes that allow for public reasoning and social choice to attend to
judgements about justice. Yet this critical role for democracy is undermined in Sen’s
elaboration in the absence of requirements of justice that would protect democracy’s
fair and effective functioning in a manner consistent with capability egalitarianism.” (S.
Srinivasan 2007, 457)
142
Thus, “Sen's concept of democracy seems an idealistic one where political power,
political economy, and struggle are absent.”(Stewart and Deneulin 2002, 64)
 147

the debate between David Crocker and Martha Nussbaum (ably defended
by Rutger Claassen) about the role of a theory of just democracy and
substantive philosophical investigation respectively for determining
which capabilities matter for social justice purposes. Section IV
contrasts two conceptions of philosophical citizenship embodied by
Nussbaum and Sen. Section V outlines and defends Sen’s ‘idea of
democracy’ as a perspective different from either form of political
moralism and with concerns orthogonal to that debate.

I. SEN’S ‘POLITICAL GAP’

As Sen has emphasised, multiple, quite different, capability lists can be
developed for different evaluative purposes. In fact he argues that such
pluralism of perspectives is required by the complex nature of the
objects we are concerned with, such as well-being or poverty. That
pluralism should be recognised as constitutive of the capability
approach – we must avoid ending up stuck with “a grand mausoleum to
one fixed and final list of capabilities” (Sen 2004a, 80).143 Thus, social
scientists may come up with various lists suited to analysing different
kinds of capabilities in different contexts.144 Sen himself has used a set
of very elementary capabilities to analyse the extent of severe poverty
(Drèze and Sen 1989; Drèze and Sen 2002), and also suggested the
relevance of a set of more complex capabilities for their instrumental
linkages for human development purposes (Sen 1999a). In all cases of
social evaluation, Sen argues that the criteria for selection and weighting
be explicit, so that the reasoning behind these choices may be subjected
to critical reflection and scrutiny by the public.145
143
“Some of the basic capabilities (with which my 1979 Tanner Lecture was particularly
concerned) will no doubt figure in every list of relevant capabilities in every society.
But the exact list to be used will have to take note of the purpose of the exercise....
Lists of capabilities have to be used for various purposes, and so long as we
understand what we are doing (and in particular that we are getting a list for a
particular reason, related to a particular assessment, evaluation, or critique), we do not
put ourselves against other lists that may be relevant or useful for other purposes.”
(Sen 2004a, 79)
144
Thus, as I have already discussed, different lists have been drawn up to study the
nature and extent of severe poverty (Alkire and Santos 2010a), and gender inequality in
western countries (Robeyns 2003). Both of these are concerned with making
evaluations of social welfare, yet neither is political in the sense of collective valuation
that this chapter is concerned with.
145
“If informed scrutiny by the public is central to any such social evaluation (as I
believe is the case), the implicit values have to be made more explicit, rather than being
shielded from scrutiny on the spurious ground that they are part of an "already
available" metric that the society can immediately use without further ado.” (Sen
1999a, 80)
148

Sen’s argument for the relevance of the concept of capability to

evaluating social and individual advantage can be seen as an outcome of
his long-term project for broadening the informational basis of social
choice to make it more fruitful and relevant to our concerns (a point I
elaborate on in section V).146 Yet this evaluative use is a distinct exercise
from determining which capabilities are of general moral significance as
a matter of social justice. That requires a stronger degree of consensus
about which capabilities are of general moral significance and fall within
the practical reach of social commitments.
The need for substantial social consensus on the requirements of
justice, rather than merely their theoretical identification and
governmental implementation, relates in part to Sen’s concern that the
idea of justice should motivate public sentiment as well as government
action (such as passing new laws). Unless these claims about justice are
generally accepted, they will not be seen as legitimate, and so they will
not be fully realised in social practice. Not only can one not assume full
compliance with the principles of justice written into one’s constitution,
but one cannot even assume full compliance with the laws themselves if
large numbers of people disagree with them. Social norms are hard to
legislate for or against.147
On the other hand, if the claims about social justice which are used
to justify a society’s institutional arrangements and programmes are
widely shared, then there is scope for public collaboration in realising
them.148 For example, if a society generally agrees that the capability to
appear in public is valuable, it takes a stance also with respect to
absences of this capability (deprivations) and the causes of such (prima
facie unfreedoms). When one looks at why certain people lack this

146
That project concerns expanding the kinds of information social choice analysis can
consider, with respect both to social states and the processes by which they are
brought about (see e.g. Sen 1999b). With regard to the former, Sen has argued that
Kenneth Arrow’s famous impossibility theorem (Arrow 1951) demonstrates that social
choice requires a broader informational base that permits interpersonal comparisons,
for example in terms of individual capability. With regard to the latter he has argued
for incorporating the character of procedures into the analysis of consequential social
states, such as in terms of their fairness and respect for individual liberty (most
famously in The Impossibility of a Paretian Liberal (Sen 1970a)).
147
For example, while South Africa’s constitution bans discrimination on the basis of
sexuality, this theoretical equality, and even basic security from targeted violence, is
far from being realised in practice.
148
As Sen notes, in addition to the “adversarial function” of democratic politics in
politicising injustices and demanding public action, “The collaboration of the public is
an indispensable ingredient of public health campaigns, literacy drives, land reforms,
famine relief operations, and other endeavours that call for cooperative efforts for
their successful completion.” (Drèze and Sen 1989, 259)
 149

particular capability (such as women in Saudi Arabia), it may become

clear that specific social norms, rather than the cost of clothes or
transport, is the cause of this deprivation. Because of the underlying
agreement about the importance of appearing in public without shame,
those social norms can be more readily recognised as invidious.149
Not all capabilities that individuals have reason to value may be
suitable objects for such consideration. For example, the reasonableness
of an individual valuing professional athletic or academic prowess does
not necessarily make it an appropriate object for society to guarantee to
everyone as a demand of justice. This relates both to the possibility for
reasonable disagreement about the importance of such kinds of lives as
well as the practical impossibility of guaranteeing positional goods
(success) to all.150 There may be more agreement on guaranteeing
everyone access to a basic level of academic and athletic training. Yet
even this may be subject to feasibility constraints on the resources that
a society is willing to make available, leading to trade-offs in an
education system between physical and academic education, for
example.
As well as such narrowing of objects, the consensus approach places
constraints on what kinds of action may be taken. This relates to the
multiple dimensions that may be relevant in thinking about social
justice, and which may be in tension with each other.151 For example,
even within the idea of equality there may be multiple divergent and
appealing moral principles to consider and reconcile, such as around the
ideas of fairness and equality of opportunity. One example that Sen
gives relates to the generally shared view that life is an important
capability. It seems that, for mainly biological reasons, women tend to

149
Though social norms are hard to legislate for or against, they are thus not beyond
the reach of social choices about justice. Hence Sen’s careful analysis of the concept of
human rights as “significant ethical claims” bearing on all those “who are in a position
to help” and thus as going beyond what can be legislated for, including, for example,
equality of respect in marriage or being treated with civility in public discourse (Sen
2009a, chap. 17).
150
The standard example here comes from Scanlon: “The fact that someone would be
willing to forego a decent diet in order to build a monument to his god does not mean
that his claims to others for aid has the same strength as a claim for aid in obtaining
enough to eat” (Scanlon 1975, 659–60). Debra Satz has also argued for dropping
idiosyncratically valued capabilities from consideration in favour of more generic ones
that can be publicly justified as the basis for policies (Satz 2012, 290–2).
151
As Sen notes, “Capability is, in fact, no more than a perspective in terms of which
the advantages and disadvantages of a person can be reasonably assessed.... But
neither justice, nor political or moral evaluation, can be concerned only with the
overall opportunities and advantages of individuals in a society” (Sen 2009a, 296–7)
150

live several years longer than men. An equal opportunity view may
therefore suggest that medical care should be preferentially directed to
the weaker sex to equalise their effective freedom to live a long life. Yet
that course of action conflicts directly with intuitions about procedural
fairness (Sen 2002c; Sen 2006a). Reconciling these multiple and
sometimes contradictory principles of justice for practical purposes is,
Sen argues, a task for public reasoning rather than moral theory (Gaus
2012, 245–55).
Yet, although Sen seems to rely on some idea of deliberative
democratic politics in using the concept of capabilities in thinking about
social justice, he has not presented a recognisable theoretical account of
how that should work.152 Hence the appearance of a ‘political gap’ in his
account. Sen has made many remarks on the issue that seem to rule out
certain ways of proceeding, notably, against allowing philosophical or
technical analysis to play a determining role in valuational exercises. Yet
many critics have found his remarks about what kind of democratic
procedure he has in mind unsatisfactory. On the one hand Sen seems to
link the legitimacy of a capability valuation to an exercise in social
choice, in which valuable functionings are identified and ranges of
weights that reflect the degree of agreement in their value are assigned
to them (Sen 1985c, 40). But on the other hand he himself points out the
limitations of such exercises in the absence of substantial uniformity of
judgements (Sen 1985b, 30). He therefore points to public reason and
democratic politics as the best way to come to legitimate decisions, but
he doesn’t give a very clear account of how this ‘government by
discussion’ should work. He also makes clear that democratic political
procedures do not by themselves guarantee good decisions on valuation
- they may still merit criticism from outside the society concerned (and
indeed inside – a fundamental feature of a functioning democracy).153
Many scholars have demanded a clearer account of the proper roles
and importance of philosophical reflection, technical analysis, and
(democratic) politics in determining the content and interpretation of a
list of capabilities for social justice. In the following I discuss alternative
answers which critics have given, through a review of the slightly fierce
152
It is this suggestion that Crocker sees himself as completing: “although Sen opens
the door to an explicit engagement between the capability approach and deliberative
democracy, he has only begun to venture through it (Crocker 2009, 308).”
153
For example, Sen often criticises even robust and wealthy democracies for their
lacunas, such as the political acceptance of high levels of ‘structural’ unemployment in
Europe or scandalous early mortality rates among some demographic groups in the
USA (e.g. Sen 1999a, chap. 1).
 151

debate between Crocker and Nussbaum as representatives of the two

wings of political moralism. My aim is to show that despite being clearer
and more precise than Sen has been, both accounts have significant
practical and theoretical deficiencies that relate to their deliberate self-
distancing from politics. With this achieved, I can return to Sen’s
account and introduce my interpretation of Sen’s ‘idea of democracy’
and show that it presents a genuine and substantive alternative to
orthodox thinking about justice in political philosophy.

II. POLITICAL MORALISM

Sen’s invocation of the idea of public reasoning has been criticised from
two different directions. In the first place, Sen’s claim may be said to be
descriptively inadequate. In the real world, politics (even in
‘democracies’) appears to be mainly about power, interests and struggle,
not public reasoning. This critique focuses on power relations, broadly
construed, and the relative powerlessness of the most oppressed, and
hence the practical irrelevance of such idealised concepts as public
reasoning. In the second place, Sen’s claim may be said to be
normatively inadequate. Insofar as he doesn’t specify what the idea of
public reasoning would require in institutional terms – such as equality
of voice - it fails to provide guidance for recognising it or bringing it
about.
The two critiques are aligned in the sense that both conclude that
the choice of capabilities for social justice purposes cannot be left to
ordinary (‘democratic’) political processes. Politics as we know it – a
domain of competitive struggle between interests mediated by power -
should be bypassed entirely. Yet there are two ways in which this is
envisaged, which relate to the orientation of the original critique
towards seeing democratic politics as ‘half-empty’ or ‘half-full’. Those
who take the more pessimistic line about the nature of politics tend to
support the external determination and imposition of the requirements
of justice on the society concerned. Those who take the more optimistic
line are concerned with improving the character of the political process
itself so that the conditions for true public reasoning are realised.
This bears a striking resemblance to the schema of ‘political
moralism’ laid out by Bernard Williams, for example in his essay Realism
152

and Moralism in Political Theory (Williams 2007).154 Williams suggests

that there are two rough models for political theory. First the
“enactment model”, in which “political theory formulates principles,
concepts, ideals, and values; and politics (so far as it does what the
theory wants) seeks to express these in political action, through
persuasion, the use of power, and so forth” (Williams 2007, 1). Williams’
paradigm of this is Utilitarianism, with its panoptical external view. (I
argue that this is the model underlying Nussbaum’s capability approach
to social justice.155) Second, the “structural model”, where “theory lays
down moral conditions of co-existence under power, conditions in which
power can be justly exercised” (Williams 2007, 1). Williams’ paradigm
here is Rawls’ Theory of Justice, with its focus, outside the difference
principle, on a fair institutional structure. (I argue that this is the model
underling David Crocker’s deliberative democracy approach to justice in
terms of capability.156)
Williams argues that both models assume “the priority of the moral
over the political” (Williams 2007, 2), what Bonnie Honig has elsewhere
called that “mysterious phenomenon: the displacement of politics in
political theory” (Honig 1993, 2).157 The enactment model treats politics
as the instrument of the moral. That is, moral ends are determined
outside the political domain, and politics is considered, if it is
considered, only in strategic terms. The structural model sets (often
tight) constraints on what politics can legitimately do. That has the
effect of substantially predetermining the kind of outcome that politics
can lead to.
In the following section I discuss the debate between the two
theoretical approaches, through an inspection of the debate between
Crocker and Nussbaum (and Nussbaum’s defender, Rutger Claassen)

154
Though Sen and Williams were close colleagues, I cannot say whether Sen has been
directly influenced by Williams very late work on political philosophy as he certainly
was by Williams’ earlier writing on ethics.
155
Thus Nussbaum notes for example that “When we think about violence against
women, we see that democratic deliberation has done a bad job so far with this
problem” (Nussbaum 2005, 179). And she says that her approach “is recommended as
a good idea to politicians in India or any other nation who want to make it the basis of
national or local policy” (Nussbaum 2000, 104)
156
Thus Crocker argues that Sen needs to go further than endorsing democratic
practice: “although it is true that deliberative politics has an important role in the
“practice” of democracy, the theory of deliberative democracy can enrich the ideals of
democracy, shape new institutional devices, and guide citizens in the practice of
democratic deliberation”(Crocker 2009, 308).
157
Honig argues though that there are exceptions to this tendency, such as Hannah
Arendt and Friedrich Nietzsche.
 153

around the central question of the relationship between democracy,

philosophers, and justice. As will be clear, both these theoretical
perspectives are informed by and try to accommodate to some extent
the nature of politics. They are not the work of ‘armchair theorists’, and
thus the distinction between them and Sen’s approach is not categorical
but rather a matter of emphasis. Nevertheless, both distance their
projects for social justice from politics as we know it.

III. DEMOCRATIC OR PHILOSOPHICAL THEORY? CLAASSEN VS.

CROCKER VS. NUSSBAUM
Let us begin with Crocker’s critique of Nussbaum and proposal of a
(version of) the theory of deliberative democracy to address the
question of capability list-making for social justice purposes. In chapter
6 of his Ethics of Global Development, Crocker criticises Nussbaum’s
approach to the valuation question on the grounds that she neglects the
political dimension almost entirely in her account of how her ideal
constitution is to be arrived at, interpreted, implemented, or reformed
(Crocker 2009, 199).158
As a result, Nussbaum misses the ‘agency’ aspect of democratic
participation.159 Citizens as agents have an interest in collectively
deliberating about and deciding on the nature of the society they want
to live in. Crocker argues that Nussbaum’s efforts are directed at the
constitutional level, rather than the democratic process, because she
worries about the danger of democratic majoritarian tyranny (for
example, with respect to women’s capabilities and voice, or those of
religious minorities), and because she believes that democratic politics
may result in trade-offs between the more and less popular items on her
list of incommensurable and equally essential central capabilities
(Crocker 2009, 204–5). As a result, Nussbaum has very little to say about

158
A further point, not mentioned by Crocker, is that Nussbaum appears to have an
inordinate faith in government as both benevolent and the only agent of justice. In
practice this is conceived even more narrowly, in terms of writing progressive
constitutions to be interpreted by progressively minded supreme courts. India, on
which she has written extensively, already has both. Yet they have not been enough to
make up for the extensive failings of Indian politics and public administration, which
Sen has discussed and criticised in some detail (e.g. Drèze and Sen 2002). One wonders
therefore what promulgating a new even better constitution incorporating Nussbaum’s
recommendations might in itself achieve. I will come back to this point below.
159
Recall Nussbaum’s view of social choice: “A habituated preference not to have any
one of the items on the list (political liberties, literacy, equal political rights, or
whatever) will not count in the social choice function, and an equally habituated
preference to have such things will count” (Nussbaum 2000, 149).
154

how the agency of citizens will play a role in the achievement of

minimal social justice, which is the ultimate goal of her capabilities
theory.160
Crocker also notes that Nussbaum’s focus on the political
relationship between individuals and government (in terms of individual
rights and constitutions) comes at the cost of neglecting the political
relationship between citizens.161 As a result, she misses the ‘constructive’
value of democratic deliberation (identified by Sen, in addition to its
more obvious intrinsic and instrumental values).162 Open public
deliberation is important for people’s conceptualisation and
comprehension of both their own individual needs and social standards.
(I will come back to this important point below in elaborating Sen’s
“mausoleum” argument against a fixed list.)
Crocker proposes a theory of deliberative democracy that he believes
addresses these issues, building on Rawls’ influential definition163 and its
theoretical development, particularly by Henry Richardson and Archon
Fung & Erick Olin Wright. It attempts to square two core values.
Development should be democratic – decided by the people concerned -
but also just.164 Therefore, development requires a theory of what counts
as a legitimate (just) democratic process. Crocker argues that he is

160
For example, although Nussbaum makes much of the Rawlsian idea of an
‘overlapping consensus’ in justifying her capability list as a partial theory of justice
compatible with political liberalism, she acknowledges that the socio-political
conditions required for her list to be the object of an actual overlapping consensus are
not yet satisfied (Nussbaum 2011a, 89–93). Thus the role of the overlapping consensus
is purely hypothetical and justificatory. It is not part of how Nussbaum expects her
theory of justice to be realized, which is by the leadership of enlightened politicians
and international civil servants rather than the demands of popular politics.
161
In contrast to Sen, for whom “the central issues in a broader understanding of
democracy are political participation, dialogue and public interaction (Sen 2009a,
326).”
162
“Open discussion, debate, criticism, and dissent....are crucial to the formation of
values and priorities, and we cannot, in general, take preferences as given
independently of public discussion” (Sen 1999a, 153 emphasis added)
163
“The definitive idea for deliberative democracy is the idea of deliberation itself.
When citizens deliberate, they exchange views and debate their supporting reasons
concerning public political questions. They suppose that their political opinions may
be revised by discussion with other citizens; and therefore these opinions are not
simply a fixed outcome of their existing private or nonpolitical interests. It is at this
point that public reason is crucial, for it characterizes such citizens’ reasoning
concerning constitutional essentials and matters of basic justice.” (Rawls 2003b, 138–
9)
164
Crocker’s central concern is with development, though, like Sen, he takes an
expansive view of this to include confronting “urgent human problems” wherever they
may be found and not only in the so called developing world. His approach is thus
compatible with the social justice concern of this chapter.
 155

merely developing Sen’s own positive endorsement of the practice of

deliberative democracy by specifying what a suitable theory of
deliberative democracy would require.165 Once specified, deliberative
democracy can be evaluated and worked towards in the real world. Its
requirements, for example, include rough equality and passing a
threshold of sufficiency for all adult citizens in the dimensions of
political liberty, the rule of law, and economic functioning, as well as
procedural fairness in the deliberative mechanism itself (Crocker 2009,
317–9).
In the course of defending substantive capability lists like
Nussbaum’s, and elaborating on the proper relationship between the
philosopher and the polis, Claassen delivers a fairly comprehensive
refutation of Crocker’s argument. Claassen notes that the democracy
theorist’s approach is justified in part by a “political” concern that
legitimate decisions about the metric and rule of social justice can only
be made by the people concerned, and only democratically. It should not
be determined and imposed by a “philosopher king”. Yet, rather
obviously, the democracy theorist has not escaped the problem he sets
himself. For if the justification for this exercise is that it is illegitimate
to impose an externally derived and justified substantive account of
justice on a society, then the ‘theory’ of just democracy must fall to the
same problem. As Claassen puts it:

If the democratic position holds that only a democratic process can

deliver the requisite legitimacy, then it needs to answer the question
whether any democratic process will do; or, to put it differently,
what it means for a process to be (sufficiently) democratic. This will
lead the democratic position to develop a (or endorse an already
existing) theory of democracy, e.g. one of the theories of deliberative
democracy ....This means that the philosophical modesty at one
point (in the theory of justice) requires philosophical outspokenness
at another point (in democratic theory) (Claassen 2011, 498)

The further problem concerns the substantive perquisites of just

democracy (Claassen 2011, 498–9). When identified in capability terms,
as capability theorists such as Crocker do, this specifies a great deal
about which capabilities matter and how they are to be distributed

165
For example, Crocker notes that the ‘social choice exercise’ promoted by Sen is an
unorthodox interpretation of social choice theory, since it includes a deliberative
exercise which takes it beyond the mere aggregation of voting information.
156

(generally, equally).166 The capability approach democrat’s theory thus

imposes quite significant decisions about what capabilities justice
requires before any legitimate democratic deliberation can take place,
and it does so for the purported reason that only the people concerned
can legitimately decide what social justice requires. As Claassen points
out, this is circular.167
Claassen goes on to distinguish the “philosopher citizen”,168 who
develops and tests philosophical accounts of justice and proposes them
for public consideration and deliberation, from the “philosopher king”
The philosopher king considers the practical legitimacy of his theory to
derive entirely from its claim to philosophical truth, and considers that
citizens should conform to it because of that truth. In contrast, the
philosopher citizen offers his account as an input to the political
process, and sees its practical legitimacy as dependent on its success in
that process rather than as deriving from its essential truth (Claassen
2011, 501). Claassen argues that Martha Nussbaum should be seen as a
philosopher citizen because she explicitly structures her account as a
proposal for democratic consideration (the distinction between the
theoretical justification for her theory and its local political
implementation noted in Ch2.II). He argues that the philosopher qua
citizen faces no legitimacy problem in putting forward her best account
of what justice requires.
Claassen then argues for moderating the “epistemic” justification for
the democratic theory approach – that “The man who wears the shoe
knows best that it pinches and where it pinches, even if the expert
shoemaker is the best judge of how the trouble is to be remedied”. 169 He
argues first that philosopher citizens provide a service to democracy by
providing fully worked out theories for their scrutiny and consideration.
In this sense it is the philosopher, not the democratic theorist, who is
more respectful of democracy. Second, he argues that the philosopher
citizen can (and should) become a “philosopher investigator”, who
crosses the boundary between philosophical theory and real world data
(on the model of Jonathan Wolff and Avner de-Shalit’s empirical
application of Nussbaum’s account (Wolff and De-Shalit 2007)).
Claassen succeeds, I believe, in defeating Crocker’s case for the
priority of democratic theory. Yet this does not mean that we should

166
(cf also Bohman 1997; Knight and Johnson 1997)
167
This circularity problem applies generally to political egalitarianism (Peter 2006).
168
A term he borrows from Michael Walzer (Walzer 1981)
169
As Crocker quotes John Dewey (Dewey 1954, 207–8)
 157

accept his conclusion of adopting the “philosophical solution” of

establishing the list of capabilities for justice “as a matter of
philosophical theory” (Claassen 2011, abstract). Or rather, we must be
more precise about what we mean by the philosophical solution.
Claassen diagnoses the distinction between the democratic and
philosophical positions as gradual rather than categorical (Claassen
2011, 497). If Nussbaum counts as a philosophical citizen, and even a
philosophical investigator for her incorporation of empirical research
and cross-cultural dialogue, so too must Sen. Yet, while Sen merely
proposes that capability be the metric in which advantage is considered
for social justice purposes, Nussbaum clearly proposes a far more
complete, far more theoretical account (while admittedly allowing for
local specifications). The question I turn to now is whether this
gradation matters, both for the democratic legitimacy and the
democratic feasibility of a capability account of social justice.

IV. THE ROLE OF THE PHILOSOPHER CITIZEN: SEN’S CASE AGAINST

LISTS
In this section I will consider Sen’s objections to lists, going beyond
Crocker’s discussion. Before doing so it is important to emphasise that
Sen is concerned with a quite different problem than that considered by
political theorists such as Nussbaum, Crocker, and Claassen. His
concern is not to identify what a just society would look like, and
therefore he need not choose between either the enactment model
(Nussbaum; Claassen) or the structural model route (Crocker). Rather his
concern is with mobilising public action against gross injustices. This
means that his method is quite different. For example, rather than
claiming to speak for the conscience of humanity, as political moralists
tend to do, he seems to consider himself as speaking to the conscience
of an actually existing public.170
This makes his task rather more straightforward, for one does not
need to develop a complete and consistent theory of justice to identify
whether or not child malnutrition is a bad thing, in order, on Claassen’s
account of philosophical citizenship to be able to “offer the democratic

Thus, Sen’s note in the preface to Development as Freedom that, “I have, throughout
170

my life, avoided giving advice to the "authorities." Indeed, I have never counseled any
government, preferring to place my suggestions and critiques—for what they are
worth—in the public domain.” (Sen 1999a, xiv)
158

process something to digest” (Claassen 2011, 502).171 Nor is this an issue

of deep moral disagreement, calling for a theory of deliberative
democracy involving ideals of reciprocity, publicity, and
accountability. 172
It is true that some forms of oppression, such as
around gender, caste, and race, may be socially divisive issues and hence
may be said to be characterised by broad moral disagreement. Yet, Sen
suggests, such disagreement doesn’t actually go very deep, for the
arguments for such discrimination do not survive the critical scrutiny
permitted by quite basic political freedoms and institutions, if that
opportunity is properly taken up. The Taliban, for example, doesn’t
promulgate its ‘moral’ code about women’s roles through moral
argument, but by preventing such argument.
Let us turn now to considering the difference in degree between the
two philosopher citizens, Sen and Nussbaum. Sen has explicitly refused
to endorse a single list of important capabilities not only to make sure
any list adequately relates to its specific purpose, but also to safeguard
the scope of public reasoning.

My own reluctance to join the search for such a canonical list arises
partly from my difficulty in seeing how the exact lists and weights
would be chosen without appropriate specification of the context of
their use (which could vary), but also from a disinclination to accept
any substantive diminution of the domain of public reasoning. The
framework of capabilities, as I see it, helps to clarify and illuminate
the subject matter of public reasoning, which can involve epistemic
issues (including claims of objective importance) as well as ethical
and political ones. It does not - and cannot - displace the need for
public reasoning. (Sen 2004d, 333 fn. 31 emphases added)

Claassen, however, argues the opposite case. He claims that the

philosopher citizen is actually more democratic than the democracy
theorist because she provides the public with fully developed and
academically tested claims concerning the true nature of justice about
171
Thus, when invited to give a lecture at the Indian parliament, Sen told his audience,
speaking it would seem as a philosopher citizen, that India’s success in preventing
famines (that monstrous and repeated failure of British Imperial rule) is a great
achievement of Indian democracy, but its failure to provide literacy and nutrition to its
children, on the importance of which all supposedly agree, is its great shame (Sen
2008b).
172
Amy Gutmann and Dennis Thompson proposed their influential account of
deliberative democracy to deal with deep contemporary moral disagreements within a
polity (specifically, the USA), such as over affirmative action and the legalization of
abortion, and not for what to do about deep moral agreements, such as against slavery
(Gutmann and Thompson 1996).
 159

which they may have a real debate. Democrats like Sen, who eschew
such use of philosophical theory, are unable to offer the public any
complete theories of justice, and thereby “leave the public debate
empty-handed” and “have to let citizens struggle to decide for
themselves what to think about capabilities” (Claassen 2011, 503).

The philosophical position offers the democratic process something

to digest. It enriches public debate by offering the most enlightened
theories that it can come up to. At the level of theory, of course,
there are likely to be competing philosophical theories instead of
one. So the public will quickly realize that truth is a complicated
matter in philosophical theory, or else it will be forced to realize this
by proponents of competing views in the public arena, calling upon
competing philosophical theories. It will then use philosophical
theories as clarifications, systematizations, sources of inspiration,
etc. of the positions that it wants to defend. (Claassen 2011, 502)

There is a certain plausibility to Claassen’s claim here, yet it does

not seem really credible to accept it as a description of how democracies
generally work. There have been occasions when the theoretical quality
of the public political debate has approached what Claassen seems to
assume is normal, such as in Britain over the Irish question and Lords
reform between 1886 and 1914; or, in the USA, the Philadelphia debates
over states’ rights, and the degree and nature of the democratic
franchise (of which The Federalist Papers is the finest example) (Crick
2002, 47). Yet this is hardly representative of politics as we know it. It
does not seem reasonable to presume that highly theoretical accounts of
justice are what democracies are looking to digest. It certainly does not
seem a reasonable foundation for the assertion that Sen’s more humble
contribution - the identification of the capability space as a suitable
metric for considering advantage - “leave[s] the public debate empty-
handed”.173
I don’t think the choice between the approaches of Sen and
Nussbaum rests on this, since the contribution of philosophical theories
“may trickle down to public debate” in more indirect ways, as Claassen
notes (Claassen 2011, 503). Yet I believe it does remove Claassen’s

173
Indeed, the public influence of Sen’s minimalist evaluative version of the capability
approach has in fact been much greater than Nussbaum’s theory of justice version.
‘The public’, from policy-makers to NGOs to citizen activists, seem overwhelmingly to
prefer Sen’s version to Nussbaum’s, for exactly the reasons Sen has given: its breadth
and flexibility make it a useful tool for evaluating and politicising injustices and
inequalities.
160

strongest argument in Nussbaum’s favour. We can now turn the debate

around and consider why Sen was so concerned that a single canonical
list would mean a “substantive diminution of the domain of public
reasoning”.
I think we must draw here on Sen’s account of the ‘constructive
value’ of democracy (which Crocker valued so highly): that it allows the
people concerned not only to contribute their own ideas and values
(agency), and their own understanding (their epistemic perspective
about where the shoe pinches), but also to come to new understandings
which may well be shared judgements. A constitutional canonical list
may constrain a polity’s collective imagination, which requires dissent
as well as collegiality. I think this is also behind Sen’s concern that
Nussbaum’s (then explicitly Aristotelian list) “may be tremendously
overspecified” (Sen 1993b, 47). As Sen puts it:

[T]he practice of democracy gives the citizens an opportunity to

learn from each other, and can also profoundly influence the values
and priorities of the society. Even the idea of 'needs' (including the
understanding of 'economic needs'), which is often taken to be fixed
and well-defined, can respond to public discussion and exchange of
information, views and analysis. In this sense, democracy has a
constructive importance, in addition to the intrinsic value it has in
the lives of the citizens and its instrumental role in political
decisions. Value formation is as much a democratic activity as is the
use of social values in the determination of public policy and social
response. (Drèze and Sen 2002, 25 original emphases)

Sen argues positively for the incompleteness of the social choice

approach rather than the theory founded list approach. A list, he points
out, narrows what public reasoning can consider by saying that only
those items matter.174 Whereas the constructive aspect of public
reasoning can produce unanticipated yet politically significant justice
oriented discussions around quite different values. One example Sen
often mentions is of how issues surrounding fertility have been opened
up to public discussion in many still poor countries, and specifically

174
“To decide that some capability will not figure in the list of relevant capabilities at
all amounts to putting a zero weight on that capability for every exercise, no matter
what the exercise is concerned with, and no matter what the social conditions are. This
could be very dogmatic.....(Sen 2004a, 79).”
 161

how a woman’s right to have a smaller family has become recognised as

legitimate.175
The case I consider here – India’s Mohatma Gandhi National Rural
Employment Guarantee Act (NREGA) – has an even clearer political
character. Although any links it might have to the capability approach
would be quite indirect, I believe it provides a good case study of the
potential of quite basic democratic opportunities, even in impoverished
polities with significantly flawed democratic institutions, for
articulating, debating, and acting on new and unanticipated moral
claims.
NREGA was passed in 2005 and is now being implemented in all
rural districts in India.176 Every household (or nuclear family thereof) is
guaranteed 100 days employment per year of unskilled manual labour
on public work schemes as a legal right at the state agricultural
minimum wage. Priority is given to women and work projects have to
employ a minimum quota of at least one-third women. When someone
registers for work it must be provided within 15 days or unemployment
compensation must be paid. The public works are intended to create
durable assets for the agricultural economy, particular relating to water
supply.177 Total cost for 2011, $9 billion.
175
“[I]n those parts of the so-called Third World in which there has been increased and
extensive public discussion of the consequences of frequent childbearing on the well-
being and freedom of mothers, the perception that a smaller family is a “basic need” of
women (and men too) has grown, and in this value formation a combination of
democracy, free public media, and basic education (especially female education) has
been very potent.” (Sen 1995b, 17)
176
This account is based on (Dey, Dreze, and Khera 2006; MacAuslan 2007; Green
2008). To my knowledge, Sen himself doesn’t seem to have been especially involved in
NREGA, though The Economic Times (of India) has quoted him giving it his approval:
“NREGA reaches out to the people and gives them income. It is an enhancer of
capability. It enhances their self-respect and participation in life and community”
(2009). However he has previously written on the importance of public employment
schemes in protecting and promoting entitlement to food (see e.g. Drèze and Sen 1989,
113–118). He has also talked about how the capability approach would look at the
value of employment and see rich world structural unemployment as a deprivation:
“unemployment is not merely a deficiency of income that can be made up through
transfers by the state...; it is also a source of far-reaching debilitating effects on
individual freedom, initiative, and skills...the ‘social exclusion’ of some groups
and...losses of self-reliance, self-confidence and psychological and physical health”
(Sen 1999a, 21). Sen’s long-term co-author on his capability based research in India,
Jean Drèze, was heavily involved in promoting NREGA through published research,
journalistic coverage, and also as a member of the influential National Advisory
Council, set up by Sonia Ghandi after the 2004 election (MacAuslan 2007, 4).
177
There are of course valid criticisms to be made of the structure of the programme.
In particular, there is justifiable scepticism about whether the works themselves create
many valuable assets for the rural economy (previous public works generally haven’t)
and one could suggest that simply giving money to poor people without making them
162

Public employment schemes have a dismal history in India of

corruption, incompetence, and waste. NREGA has been designed to
mitigate at least some aspects of this. It is intended to solve the problem
of identifying welfare recipients on the basis of need by building in a
self-selecting element: anyone who requests unskilled manual labour at
the minimum wage may be presumed to be in need of the income it
provides. It also creates a legal right to employment, rather than the
piecemeal schemes in the past which were often hostage to short-term
political interests. It embodies considerable attention to implementation
issues, particularly with respect to transparency and accountability.
Muster rolls, payment rates, payment processes are all supposed to be
public information; village assemblies and councils are given roles in
auditing and implementing particular work schemes. India’s often
troubled relationship between the central government and state
governments is mitigated by a clever incentive system whereby the
central government pays most employment costs, but state
governments are incentivised to carry out the organisation necessary to
implement the act since the unemployment costs fall on them if people
in their state cannot be provided employment.
The history of NREGA is somewhat complicated (MacAuslan 2007),
as is the story of most political achievements, involving activism, media
publicity, state institutions, political parties and elections, and
serendipity. Activists and organised social movements from various
backgrounds had helped to shape a prevailing civil discourse around
poverty and rural distress and were particularly encouraged by the
Indian Supreme Court (in its 2001 decision mandating mid-day meals
for schoolchildren). More formal political factors included the internal
party dynamics of the Congress Party (leaders like Sonia Ghandi
supported it) and the 2004 national elections (Congress’s manifesto was
aimed at rural voters who felt left out of increasing urban prosperity
and was perhaps particularly open-handed because they hadn’t expected
to win). There was a tipping point when NREGA came before parliament
itself. Initial drafts seen to be ‘diluted’ (by civil servants anxious about
the costs) were characterised as ‘anti-poor’ and made the subject of a

labour in the sun would be just as effective and less expensive to administer. While
this may be true, up to a point, it misses the political feasibility of guaranteed
employment rather than a handout. Supporters of the programme also believed that if
people felt that they earned their entitlements through work, they would be far more
assertive of their rights to them (the looting by Indian public officials of welfare
programmes focussing on hand-outs to the poor has previously been accompanied by
general apathy among the poor, rather than outrage).
 163

systematic campaign involving various regional and national social

movements and leftist political parties, who organised rural marches,
sit-in protests, and public hearings, with a great deal of sympathetic
media coverage. The revised version was passed unanimously by
parliament, but not without continuing criticism from all sides.
One of the interesting things about the political campaign for
NREGA was that it engaged the support of a coalition of diverse social
movements and political activists because it was seen to be
instrumentally linked to ameliorating a number of distinct but
overlapping social justice concerns. If NREGA had not been the focal
point, each of these groups might have campaigned separately for their
various causes, and perhaps only at the regional level. Or another
national programme on a quite different issue might have become the
focal point of a similar coalition.
The coalition included national organisations like the Right to Food
movement, and associated groups, which saw NREGA as an important
way to support the adequate nourishment of the poor by addressing the
rural underemployment that reduced entitlement to food or made it
insecure. Women’s movements were particularly concerned with rural
women’s lack of employment opportunities and equality within
employment. NREGA’s prioritisation of women candidates would allow
more women access to non-domestic waged employment. This was also
expected to improve the status and bargaining position of women within
households, which is substantially dependent on perceptions of
household contribution that are in turn strongly related to external
waged employment.
Left-wing political parties and labour movements backed NREGA in
part because of the long-term political opportunities it represented.
Informal labour, particularly rural, had long been under-represented
politically despite making up far more of the workforce than the formal
labour sector with its organised and politically connected unions.
Leftwing movements believed guaranteed employment would not only
increase the bargaining power of such workers with respect to other
local employers and lead to a general rise in working conditions and
pay, but also create the conditions for a more assertive rural labour
164

force who would seek a greater voice and representation in the wider
political sphere.178
Rural underemployment was also responsible for the routine mass
seasonal migration of rural labourers to cities or other states whenever
there was no agricultural waged work available at home. The economic
necessity for migration disrupted family and community life and
institutions, and led to many associated problems, such as discontinuity
in children’s education. The income security of guaranteed employment
over low seasons would make such migration less necessary.
The decentralised implementation of NREGA also appealed to
supporters of the Panchayat system (local government) since village
public assemblies (Gram Sabhas) were to be involved in the selection of
work projects and auditing, and elected village councils (Gram
Panchayats) would have a large role in its administration, setting village
development plans (work projects), and their implementation.
From the perspective of the capability approach one can see the
creation of NREGA in terms of the political valorisation of a new
‘capability for employment’.179 It is the product of a combination of
public reasoning (broadly construed) and concerted public action. As
well as its direct importance to the lives of rural citizens (its intrinsic
value), this capability for employment was additionally justified by its
instrumental linkages to supporting other important capabilities, such
as for nourishment and political representation. People who disagreed
severely about the nature of a just society could nevertheless agree that
this capability was important, and was worth making the focus of a
political struggle involving protests as well as public debates.
As I have noted earlier, an important feature of the capability
approach is that it considers individuals, even the very poor, as active
agents in their own destiny rather than as patient recipients of
dispensed benefits (Sen 1999a, xiii, 19). As well as the intrinsic value of
respecting the agency of individuals concerned, thinking in terms of
agency has instrumental value in the achievement of public actions
relating to social justice. It directs attention to issues like incentives
(understood generally, not only in terms of ‘market incentives’) and real-
178
The increased enfranchisement of the rural poor was a significant outcome of the
Maharashta state Employment Guarantee Scheme on which NREGA was modelled
(Drèze and Sen 1989, 116 fn. 29).
179
It is not anticipated for example by Nussbaum’s philosophically derived list,
(although that does include “the right to seek employment on an equal basis with
others” as part of Capability 10, Control over one’s environment, under Material aspects
(Nussbaum 2011a, 33–4)).
 165

world behavioural norms, such as the unintended but foreseeable

consequences of policy interventions. It also directs attention to how
people may be helped to become engaged in overcoming problems in
implementation, from incompetence to corruption, through both
“collaborative” and “adversarial” engagement (Drèze and Sen 1989,
259).180
The issue of enhancing political representation, particularly pressed
by the labour unions, is especially interesting because it concerns the
use of democratic political opportunities not only to demand public
action but also to enhance the democratic quality of the political regime
itself. One may note that the political agitation around the right to
employment was almost entirely organised and carried out by people
who would not be affected by it (especially, members of the urban
middle-classes and urban trade-unions), acting in ‘solidarity’ with the
nearly voiceless rural poor. As Sen has noted, the history of India shows
that such solidarity extended from the politically influential to the
politically excluded is an important but undependable method of
representing the interests of the excluded. “Self-assertion” of opinion by
society’s underdogs is also necessary, both for its intrinsic value, and
because it makes solidarity more effective and checks its failings (Drèze
and Sen 2002, 28–32). The anticipated expansion of political
representation among the rural poor due to NREGA thus exemplifies the
scope that democratic political opportunities afford.
The preceding two sections have discussed the contentious debate
between democratic and philosophical theorists over how to determine
which capabilities should be considered a matter of social justice.
Although concerned with a specific capability view of the metric of
justice, I suggested that the debate can be seen as a clash between
representatives of the two distinct wings of political moralism identified
by Williams – the structural and enactment models. The discussion
identified a significant contradiction in the justification of the structural
model (Crocker’s deliberative democracy proposal) with respect to its
claims both that the people should decide what justice requires and that
only a philosophically adequate regime of just democracy can determine
180
Bhela Bhatia and Jean Drèze summed up their critical review of the initial roll-out of
NREGA thus, “All said and done, NREGA has created a sense of hope amongst the rural
poor. This sense of hope can be further strengthened if people understand that the act
gives them employment as a matter of right, and that claiming this right is within the
realm of possibility. Translating this latent energy into organised public pressure is the
best way to ensure that the implementation problems reviewed in this article are
addressed without delay.” (Bhatia and Drèze 2006, 3202)
166

the people’s will. Yet the alternative, in which the philosopher citizen
(such as Nussbaum) works out a theory of justice and submits it to the
political process, also seems to miss something important about
democracy that goes beyond legitimacy and epistemic authority: the
constructive possibilities of democratic politics emphasised by Sen. The
following section elaborates on Sen’s positive defence of democratic
politics (i.e. beyond the epistemic and legitimacy problems one may
suspect still linger around Nussbaum’s enactment model). I call this
non-ideal, pragmatic but optimistic approach Sen’s ‘idea of democracy’
(after his Idea of Justice, with which it has an intimate connection).

V. SEN’S ‘IDEA OF DEMOCRACY’: SOCIAL CHOICE AND GOVERNMENT BY

DISCUSSION
Sen’s ‘political gap’ – his failure to rule on how politics should relate to
capability justice - is not the lacuna it is taken to be. Rather, it reflects
Sen’s belief in the idea of democracy, a belief which is grounded in his
extensive empirical, theoretical, and philosophical analysis of politics
across his career. Sen’s faith in actual democratic politics is unusual in
political theory and relates to his comparative approach to justice. It
leads him to argue for democratic politics and for a better democratic
politics, but never for overcoming politics itself.
First, Sen argues for taking the instrumental role of politics
seriously. While many moral philosophers focus on the injustices and
other failures of politics, including of more or less democratic politics,
Sen focuses on its potential - what it can achieve and how. That
potential is not only for achieving public action on remediable
injustices, but also for improving the political regime itself, for example
by extending the franchise to women or ethnic minorities, or
introducing freedom of information laws that make it easier to hold
governments to account. The challenge for those seeking to reduce
injustice is thus a political challenge - to take up, and enhance, the
political opportunities available. As Sen puts it, somewhat bluntly, “In a
democracy, people tend to get what they demand, and more crucially,
do not typically get what they do not demand (Sen 1999a, 156).”
Second, Sen sees in the process of democratic politics a productive -
transformative - dimension that cannot be replicated by moral theory.
He connects political engagement with public interaction and dialogue
about how society (and politics) should be improved. This complements
the instrumental aspect of politics because it concerns the
 167

determination of what it is that people try to demand through politics.

Sen’s account of this brings together two distinct traditions – social
choice and public reasoning – in a way which has not generally been
appreciated.181

a. An Expansive View of Social Choice

Sen has made substantial contributions to the related fields of social

choice theory and welfare economics (for which he received the
Economics Nobel Prize in 1998).182 Welfare economics and social choice
theory are generally understood as overlapping but distinct fields, the
one concerned with evaluating society’s welfare and the other with
making collective choices that properly reflect the interests and
concerns of society’s members. More specifically, welfare economics has
traditionally been concerned with the narrow question of evaluating the
welfare of a social state in terms of an aggregation of the utility
functions of its constituent individuals as the basis for assessing claims
that particular policies or events increase or reduce social welfare. Sen
considers such a narrowly defined welfare economics as a sub-field of
social choice theory, broadly conceived to include social choices as to
how to evaluate welfare (in ways that can go beyond aggregate utility
and culmination outcomes). Social choice theory, Sen argues, “provides a
general approach to the evaluation of, and choice over, alternative social
possibilities (including inter alia the assessment of social welfare,
inequality, and poverty) (Sen 1999b, 349).”
Sen notes that the traditional concerns of social choice theory
(voting) and welfare economics (aggregate utility maximisation)
substantially converged in Kenneth Arrow’s famous resurrection of the
social choice school (Arrow 1951), since preference orderings and voting
information are analytically similar (Sen 1999b, 352). This convergence
was due to the informational constraints on interpersonal comparisons
imposed by Lionel Robbins and other economists influenced by logical
positivism. Indeed much of Sen’s work in the field of social choice has
been concerned with showing that the possibility of social choice

An exception is Mozaffar Qizilbash (Qizilbash 2007)

181
182
For detailed discussions of Sen’s various contributions to economics leading up to
the prize see (Arrow 1999; Atkinson 1999; Sen 1999b). This work has also influenced
philosophers. John Rawls for example referred extensively to Sen’s book, Collective
Choice and Social Welfare (Sen 1970b), in A Theory of Justice (Rawls 1999 Revised
Edition ).
168

requires expanding the kind of information considered, with respect

both to social states (beyond utility) and the processes by which they are
brought about (such as respect for liberty) (see e.g. Sen 1999b).
Social choice theory is an extremely flexible, though thin, formal
framework that Sen argues can even be applied to moral theories.183
Nevertheless one can make a distinction between two different concerns
that social choice may have, concerning the evaluation of social welfare
and the evaluation of social judgements (cf Sen 1995b, 5).
The former can be viewed as following in the fundamental concern
of welfare economists with evaluating social welfare in terms of an
appropriate aggregation of the welfares of the individuals concerned,
while dropping the orthodox welfare economist’s restrictions on what
kind of information can be considered in evaluating individual welfare.
The capability approach is fundamentally an argument for considering
information about individuals’ capability (what they can actually be and
do) in such evaluations of social welfare.
The latter can be viewed in terms of the traditional concern of social
choice theorists with finding a workable and normatively adequate way
to translate the aggregated normative judgements of individuals about
how society should be arranged into a determinate social choice. It is
this latter aspect of social choice – how to determine society’s collective
judgement - that is relevant to political decision making.
Sen’s promotion of this aspect of social choice has two dimensions.
First he has emphasised the capacity of the very general analytical
framework of social choice theory to generate actionable conclusions
even in the context of continued disagreement. (This was the feature I
discussed in chapter 3:III.) As is apparent on opening any newspaper,
there is a great deal of disagreement in any society about the best way
of arranging matters (what achievements should be sought and by what
route), but that disagreement coexists with substantial agreement about
the worst ways. Society’s ‘choice’184 may be very clear on certain issues
(such as every child’s moral claim and legal right to a good education)

183
Sen illustrates this with respect to the problem, introduced in The Idea of Justice, of
who should have a flute: the child who made it, the one who plays best, or the one who
is poorest? Though each of these moral theories provides an independent perspective
founded on a distinctive informational basis, in considering a particular case there
may often be sufficient congruence in their reasoning to permit a definite conclusion
(Sen 2009a, 396–400).
184
Of course it would be a mistake to consider a reference to ‘a society’s choice’ to
imply that a society has some kind of “organic existence apart from that of its
individual components” (Buchanan 1954, 116).
 169

while other matters remain subject to contentious dispute (such as

whether this social commitment should be realised only through the
state education system or by giving parents transferable vouchers).185 To
the extent that a social choice system makes such agreements clear, it
can be very effective at recognising gross injustices: where social reality
falls short of those commonly agreed standards.
The social choice approach to thinking about justice thus contrasts,
Sen argues, with the ‘transcendental institutionalist’ social contract
approach. It is concerned with sorting out what we as a society should
do about the problems we have now, and not with the abstract
contemplation of what the best way to organise society might be. It
differs further from the social contract tradition, at least in Sen’s
expansive formulation, by considering people as “arbitrators” rather
than “negotiators” (e.g. Sen 2006e), and by making room for the
consideration of the voices and the interests of people from outside the
polity (e.g. Sen 2002d).
This brings in the second distinctive feature of Sen’s use of social
choice, which, as Crocker noted, contrasts dramatically with the
orthodox understanding of social choice theory. A great deal of work in
social choice focuses, naturally enough, on voting behaviour. In contrast,
Sen also emphasises the dynamic features of the extended exercise of
social choice, which democratic decision-making represents. For
example, an important challenge in orthodox social choice theory is to
develop voting procedures which correctly identify each individual’s true
preferences over social states.186 Sen is concerned not only that people’s
true preferences be counted, but also with their quality, and that the
theory of social choice adequately reflect the extended participatory
engagement that characterises democratic political systems.187

185
For example, Clark and Qizilbash have used a “super-valuationist” approach to
identify the dimensions of “core poverty” in terms of the “virtual unanimity” of
ordinary people’s judgements of the dimensions and levels of achievement required
“to get by” (Clark and Qizilbash 2008). They made use of the results from a South
African survey on the ‘Essentials of Life’: where 95% or more of the respondents
identified a dimension as essential this was taken to be virtually unanimous
agreement; in identifying minimal critical levels of achievement for each dimension the
lowest 5% of responses were discarded. Their results differed from standard measures
of the “ultra-poor”.
186
And thus, for example, to exclude ‘strategic voting’ behaviour, a condition imposed
on social choice functions by the Gibbard-Satterthwaite theorem (Gibbard 1973;
Satterthwaite 1975).
187
“It is centrally important for social choice theory to relate formal analysis to
informal and transparent examination. I have to confess that in my own case, this
combination has, in fact, been something of an obsession....Our deeply felt, real-world
170

Instead of limiting the role of social choice theory to the moment of

election and the best way to interpret a snapshot of people’s subjective
preferences, Sen takes a broader view of politics as a space in which
people discuss and criticise various views about how society should
work, and as a result often come to clarify and narrow their
disagreement and also to find many ideas so wanting (incoherent,
unfeasible, or abhorrent) that they can be ruled out altogether. In
relating to ‘the market’ individuals face a closed choice over a fixed
menu, and do not have to consider the secondary implications of their
choices on others. In contrast, democracy is a social, participatory
activity. Citizens in a democracy do not only relate to politics by
choosing the option they like best from the menu presented to them
every few years, and they do not only consider the choice as about
achieving their own interests.188 First, if people take up the opportunities
allowed by democratic freedoms and institutions (including a free
media), they will play a role in determining what will be on the menu
that is formally presented at elections. Second, democracy evokes
people’s social perspective and related preferences, which may often be
at variance with the preferences they would express in their perspective
as an individual actor (for example, people vote for taxes even though
they would rather not pay them, or to protect national parks that they
have no particular interest in ever visiting).189
Sen’s understanding of social choice thus seems to incorporate
public deliberation, a very different way of looking at democracy. Sen
asserts their fundamental compatibility, and indeed complementarity.
For example he claims that deliberation can winnow out many
(ridiculous or abhorrent) options and thus make social choice
procedures more feasible.190 These complementarities have been worked
out more systematically by John Dryzek and Christian List (Dryzek and
List 2003). They argue that public deliberation has the capacity to

concerns have to be substantively integrated with the analytical use of formal and
mathematical reasoning.” (Sen 1999b, 353)
188
The presumption, common in the public choice literature, that people’s expressed
preferences about what kind of society they would like to live in entirely reflect their
self interest (the homo economicus view) and not their beliefs about social justice is
not sustainable as an empirical claim (cf Sen 1977; Sen 1995b, 15).
189
In the same way, Nussbaum points out that even people who do not see the
relevance of some of the capabilities on her list to their own lives (such as relating to
nature) and do not wish to exercise them, may still agree that such options should be
generally available (Nussbaum 2000, 153).
190
“The actual disagreements that exist may be removed through reasoning, helped by
questioning established prejudices, vested interests and unexamined preconceptions.
Many such agreements of real significance can be reached, but this is not to claim that
every conceivable problem of social choice can be settled this way.” (Sen 2009a, 396)
 171

promote the normative conditions whose combination social choice

theory is concerned with achieving (in its famous impossibility
theorems) and conclude that, “social choice theory suggests not that
democratic decision making is impossible, but rather that democracy
must have a deliberative aspect” (Dryzek and List 2003, 1).
More specifically, Dryzek and List adduce good reasons for
supposing that what they call the informational, argumentative,
reflective and social aspects of deliberation may permit various “escape-
routes” from the impossibility results of social choice theory. For
example, they argue, in line with Sen’s assertion but in a more
systematic, technical way, that the Universal Domain condition (that no
logically possible profiles of preference orderings be excluded) may be
relaxed if deliberation induces preference structuration. One way in
which this might be brought about by deliberation is by inducing meta-
level agreements about the dimension along which to conceptualise the
alternatives, for example in terms of fairness or efficiency (Dryzek and
List 2003, 12–22). That is, even where substantive disagreement about
how to rank alternatives persists, an agreement about what it is that one
is disagreeing about imposes sufficient order on the profiles of
preference orderings to reduce the threat of cycles in social preferences
and strategic manipulation and so permits democratic social choice
procedures to go through. Dryzek and List argue that this is both a
reasonable supposition to make, and is corroborated by empirical
studies of the results of James Fishkin’s Deliberative Polling® (List et al.
2006).191 Their analysis leads them to conclude that,

deliberative democracy and social choice theory are mutually

supportive. The former is concerned with identification of the
functions that deliberation ought to, and indeed can, perform in
democratic decision making, and the latter is concerned with the
clarification of the logical properties of available procedures for
solving the aggregation aspects of democratic decision problems.
Thus social choice theory shows exactly what deliberation must
accomplish in order to render collective decision making tractable
and meaningful, suggesting that democracy must in the end have a
deliberative aspect. (Dryzek and List 2003, 28)

191
Fishkin’s Deliberative Polling® has 3 stages (Fishkin 1991). First a representative
group of individuals is polled on a particular issue. Then there is a deliberative
exercise in which the participants are brought together for a couple of days to discuss
the issue, with experts and briefing materials. Then the initial poll is repeated and
changes to the answers given noted.
172

Dryzek and List’s analysis supports the feasibility of Sen’s

unorthodox concern with the exercise of social choice, that abstract
social choice analysis adequately incorporate the transformative
character of actual political participation and engagement for beliefs
and voting behaviour. That same concern with practise is also central to
Sen’s direct discussion of democratic politics, to which I now turn.

b. Democracy as Government by Discussion

Sen has paid a great deal of empirical and theoretical attention to

democracy and politics, within and beyond his work on the capability
approach and development.192 Here I will bring together the salient
features of Sen’s analysis to show how his conclusions about the
“intimate connection between justice and democracy” (Sen 2009a, 326)
are warranted. Sen is acutely conscious of the non-ideal character of
actually existing democratic systems of social choice. Nevertheless he
argues that they permit considerable scope for public reasoning about
social justice that can reach definite conclusions. He also argues that the
solution to the problems of democracy is more democracy (Sen 2003b).
Addressing the flaws in a society’s political regime is a practical political
project rather than a politico-moral theory project. That is, the failures
in a political system that hinder or undermine the collective capability
of a society for public reasoning should themselves be politicised. Thus,
pace proponents of the structural model of political moralism, Sen
argues that “A country does not have to be deemed fit for democracy;
rather, it has to become fit through democracy (Sen 1999c, 4).”
Hardnosed political realists like Ian Shapiro have accused Sen of a
certain naiveté in depicting politics in terms of the idea of public reason
rather than a Hobbesian struggle between interests over power.193 Yet
Sen’s idealism seems better understood in terms of an optimistic
reading of the facts than ivory tower fantasy.194 For example, his book
with Jean Drèze India: Development and Participation (Drèze and Sen

192
Sen’s analysis of politics is perhaps most extensive in (Sen 1999a; Drèze and Sen
2002; Sen 2009a).
193
“Sen characterizes democracy as a system of public reason and discussion. The
image he seems to have in mind is an academic seminar writ large, where the best
argument wins.” (Shapiro 2011, 1259)
194
See, for example, Erin Kelly’s defence of Sen’s endorsement of partial and
imperfectly realised public reason, between Rawlsian idealism and Hobbesian realism:
“[Sen’s] characterization of the content of public reasoning is loose. Yet it does not
describe a mere modus vivendi.” (Kelly 2012, 306)
 173

2002 particularly ch. 10) analyses the working of Indian democracy in

some detail and nuance. It introduces a distinction between democratic
ideals, institutions, and practise, and argues that India has well
entrenched democratic ideals, and institutions that function pretty
much as they are supposed to, but particularly problematic practise.195
That is, the capacity of India’s political rights, freedoms, and institutions
are not being used as effectively as they could be, either to propose
positive justice enhancing ideas or to criticise iniquitous laws, policies
and institutions.
Sen is particularly concerned that the potential of India’s democracy
is not being sufficiently brought to bear on chronic problems of poverty
to which people have become inured.196 But he views this not as a failure
of the system’s legitimacy but a problem of inadequate functioning
requiring a thoughtful scrutiny of the multiple causal factors and their
relations, such as the lack of informed public engagement (despite very
respectable voter turnouts at elections) and socio-economic inequality.
Indeed, the same critical analysis identifies grounds for optimism about
what Indian democracy can achieve. For example, Sen and Drèze identify
the evolution of social movements around new issues, which have for
example succeeded in politicising government corruption in India and
transforming public attitudes towards it (Drèze and Sen 2002, 351–2).
While noting the many flaws (especially extensive corruption) of India’s
extensive panchayat system of village-level local democracy and
governance, Sen and Drèze also note their achievements where
circumstances are conducive to their success (such as in the relatively
well educated and egalitarian state of Kerala). Local democracy also
supports democracy in general by providing a space for “learning by
doing”, with regard for example to discussing social needs and problems
or the idea of holding officials accountable for public services (Drèze
and Sen 2002, 361–3).
The instrumental value of democracy lies in its potential for
promoting sensible policies and holding public officials accountable. Sen
notes for example that no democracy has ever experienced a substantial

195
This line of analysis reflects Sen’s general scepticism about the sufficiency of
institutional accounts (whether of justice in general or democracy in particular) which
neglect consideration of the full social realisation of the principles they embody.
196
“India's success in eradicating famines is not matched by that in eliminating regular
undernutrition, or curing persistent illiteracy, or inequalities in gender relations.....
While the plight of famine victims is easy to politicize, these other deprivations call for
deeper analysis and more effective use of communication and political participation—
in short, fuller practice of democracy.” (Sen 1999a, 154)
174

famine, despite the fact that famines rarely directly affect (starve) above
5% of the population. They are perceived as a moral outrage by society
as a whole, in solidarity with the victims, and that sense of outrage is
transmitted effectively by the political system to the government
officials with the power to prevent famine (which is actually rather
simple). This feature explains both why India’s series of dreadful
famines ceased upon independence from Britain, and also why more
complicated ‘less urgent’ problems in India such as astonishing rates of
childhood malnutrition and illiteracy are permitted to continue: they are
moral outrages that have not yet been politicized as effectively as
famines.

Now take some cases of lesser success – and even failure. In general,
Indian democracy has been far less effective in dealing with
problems of chronic deprivation and continuing inequity with
adequate urgency, compared with the extreme threats of famines
and other emergencies. Democratic institutions can help to create
opportunities for the opposition to demand – and press for –
sufficiently strong policy response even when the problem is chronic
and has had a long history, rather than being acute and sudden (as
in the case of famines). The weakness of Indian social policies on
school education, basic health care, elementary nutrition, essential
land reform, and equal treatment of women reflects, at least partly,
the deficiencies of politically engaged public reasoning and the reach
of political pressure. (Sen 2008b Hiren Mukherjee memorial lecture
at the Indian parliament)

The important point here is that although we are in a second best

setting for democratic politics we should not lose sight of the real
improvements that can be achieved within that system with respect to
advancing both social justice and public reasoning about it. 197 We should
not make the mistake of conflating the actual achievements of a
democratic politics with the limits of what it can achieve. Nor should we
assume that achieving more from politics is a primarily theoretical
challenge (to design a system that will automatically produce morally

197
The state of Kerala, for example, whose residents are now among the most
economically equal, longest lived, best educated and most adequately politically served
in India, was relatively recently one of the most caste-hierarchical parts of India. Not
only was the institution of untouchability strictly observed, but lower and non-caste
residents were also subject to the intricate and demanding strictures of
“unapproachability” and even “unseeability” (Lelyveld 2011, chap. 7). Its achievements
relate not only to fortunate circumstances but to “a firm history of radical politics”
(Sen 2003a, 324).
 175

just outcomes). Improving the practise of democracy is a practical

political challenge. For example, Sen argues that the development and
reinforcement of “countervailing institutions” can amplify the voices of
the politically marginalised (Drèze and Sen 2002, 371). Such
countervailing institutions, though they may themselves be very
imperfect (Sen gives the example of trade unions in India), can provide
functional checks and balances for holding public officials to account,
and raise the profile of issues that are currently absent from the
political agenda (such as chronic hunger or unemployment, as opposed
to sensational famines), and even directly influence norms and values. 198
By moderating asymmetries of power they can help mitigate socio-
economic inequalities in access to democracy, and even the systemic
corruption of public institutions that undermines its working.
The underlying position Sen defends is of the importance of the
space between the extremes of the full realisation of our moral ideals
and their complete absence. This is the space in which we actually live,
and in which we are concerned to improve our position. I think Sen’s
justification for focusing on this space has two pillars. First, the
negative critique of the redundancy of the kind of ideal theorising that
characterises transcendental institutionalism (discussed in chapter 3:
III). This carries over from Sen’s general discussion of the idea of justice
to the idea of democracy. The transcendental institutionalist approach
focuses on the relationship between democratic ideals and institutions,
at the expense of attending to democratic practice and the scope for its
improvement. Second, the positive defence of the capacity of
imperfectly institutionalised democratic ideals to provide a partial space
for democratic practise, and the scope of that democratic practise in

198
Note that international institutions, such as Human Rights Watch or the Mo Ibrahim
Foundation, can also have this function.
Sen has himself founded two such ‘countervailing institutions’ (with his Nobel
Prize money). The Pratichi Trusts focus on primary education and health care (India)
and the social progress of girls (Bangladesh). They issue reports on systematic
problems in the public provision of basic services (Rana, Rafique, and Sengupta 2002;
Pratichi Trust 2005a; Pratichi Trust 2005b; Pratichi Trust 2009; Rana and Chakraborty
2009) and promote improvements through media coverage (e.g. Sen 2009b; Sen 2009d;
Sen 2009c), political lobbying, and collaborative engagement with the actors concerned
(such as parents, village-councils, teaching unions and state and central government
departments of education and legislatures). Their contribution to the practise of
democratic politics seems threefold. First, conducting original empirical research to
establish the facts of the matter; second, publicising their findings to the general
public; third, in co-ordination with various other social movements and NGOs, direct
engagement – both adversarial and collaborative – with the actors concerned.
176

turn to effect real improvements in social justice and democratic

practise itself.
Erin Kelly provides a substantial defence and elaboration of Sen’s
contribution in these terms (Kelly 2012). She combines Sen’s analysis of
public reason with the capability approach to evaluation and argues that
“understanding public reason as a collective capability” helps us clarify
its demands and resist scepticism about the ethical accountability of
politics and cynicism about its ethical content (Kelly 2012, 295).
Considered in terms of the capability of a society, the imperfect
realisation of public reason can be evaluated, and the various resources
and conditions that support or undermine it can be identified - and
themselves politicised.
Kelly argues that the imperfect realisation of the ideal of public
reason can have substantial effects on the form of politics. It not only
increases the ability of the people to effectively demand improvements
in social justice through politics. The requirement for public
justification, even if only half-heartedly adhered to by many parties, also
raises and broadens the standards for political legitimacy (Kelly 2012,
306). For example, Kelly characterises the spread of human rights norms
internationally since the mid-20th century as an exercise of the collective
capability for public reason, with respect to a particular subject, by an
extended international community of NGOs, activists, lawyers,
academics, and journalists (Kelly 2012, sec. VI). This evolving discourse
has severely constrained even quite authoritarian regimes, which have
been placed in the awkward position of still wanting to torture people
but being embarrassed of the public condemnation of being found out
by the many organisations, like Human Rights Watch, who scrutinise
their practices. Despite the primacy of national sovereignty in
international law, it has also led to increasingly interventionist
international norms against gross abuses (as Muammar Gaddafi
discovered with UN Security Council Resolution 1973). Kelly also notes
the role of public justification in the evolution of norms about gay
marriage in the United States. American courts have found that laws
discriminating against this group on the basis of “private” moral views
are unconstitutional, and this attitude seems increasingly widely shared.
As Kelly interprets it, opponents of gay marriage are being increasingly
challenged to provide a public justification for their opposition, and are
losing the political battle because of the general perception of the
weakness of their public reasons.
 177

Kelly’s analysis, positive proposal, and case studies, demonstrate the

significance of imperfect public reason in the real world, without
needing to resort to either model of political moralism. Kelly’s proposal
thus supports Sen’s claim that “the defects of democracy demand more
democracy”:

The value of public reasoning applies to reasoning about democracy

itself. It is good that the practices of democracy have been sharply
scrutinized in the literature on world affairs, for there are
identifiable deficiencies in the performance of many countries that
have the standard democratic institutions. Not only is public
discussion of these deficiencies an effective means of trying to
remedy them, but this is exactly how democracy in the form of
public reasoning is meant to function. In this sense, the defects of
democracy demand more democracy, not less. (Sen 2003b)

Sen’s ‘idea of democracy’ recognises the priority of politics over

moral philosophy in the sense that the legitimacy of political outcomes
or regimes, even those with democratic features, is itself a political issue
distinct from the requirements of moral theory. However, this is
consistent with his belief that politics can be tremendously improved,
and that such improvement extends to the moral standards by which
people judge the legitimacy of political outcomes and the regime itself.
Indeed ‘Politicise!’ might be Sen’s slogan for how to deal with the
shortcomings of actual political systems. It is only by politicising
corruption or the lack of women’s rights - perhaps through such non-
ideally deliberative democratic methods as public demonstrations and
civil disobedience – that those moral issues are raised as challenges to
the legitimacy of that politics.
While it may seem dispiriting to accept that improvements in social
justice depend on a mode of interaction as inherently imperfect as
politics, Sen suggests reasons for optimism. The potential for using
politics to advance justice is greater than we often assume, and so is its
scope. To the extent that the politicisation of injustice succeeds (such as
by America’s civil rights movement), it changes the rules of the society’s
politics in terms of what constitutes legitimate kinds of policies and
how they may be publicly justified. And the further that progresses, the
more the practical form of political legitimacy will approach the moral
ideals that have been generally accepted by that society. Such
politicisation can thus not only remove particular injustices, but also
make the rules of the game less unjust, one part at a time.
178

CONCLUSION
I began this chapter by laying out a challenge to the contribution of
Sen’s capability approach to thinking about social justice. There is, as
many have pointed out, a ‘political gap’ in his work about how justice in
the space of capabilities should be implemented. Two alternative
philosophical accounts (from each wing of the ‘political moralism’
approach) were advanced as competitors to Sen. Yet each were shown to
have significant theoretical and practical problems: both had a ‘political
gap’ of their own resulting from their deliberate self-distancing from the
practise of democracy.
I then turned to considering the positive contribution of Sen’s
approach. Firstly in terms of the theoretical account he does provide:
the thin but highly flexible framework of social choice. And secondly in
terms of his embrace of the practise of democracy as a space with
underappreciated potential for advancing social justice. With regard to
the former, Sen complements the orthodox concern of social choice
theory with voting behaviour with a concern for the exercise of public
deliberation that actual social choice involves. With regard to the latter,
Sen presents an optimistic but pragmatic view of the instrumental
capacity and scope of democratic practise for achieving improvements
to social justice. This particularly relates to the mechanism of
‘politicisation’, whereby social injustices, and even failings of the
democratic regime itself, are made the objects of political mobilisation.
Let me conclude by trying to answer the disappointment the political
philosopher may feel at this retreat from the rich resources of rigorous
moral theory to the flawed and unreliable space of democratic practise.
It is sometimes asserted that we should assess the legitimacy of a
society’s political regime in terms of whether they fulfil the
requirements of our best account of justice. Yet, in actuality the
practical legitimacy of a social order depends on the perception of those
subject to it, a perception that certainly evolves but cannot be
independently identified with an abstract and moralised definition, of
liberalism say, unless that is itself taken up by the members of that
society (Williams 2007). It is the practise of democracy, as promoted by
Sen, that presents the opportunity for a society’s sense of what a
morally legitimate social order requires to be put in play, for the
distance between our ideals and political reality to be identified and
challenged. There is much scope here for the moral philosopher, qua
 179

philosopher citizen, to participate in that slow and piecemeal process of

improving the world we live in by contributing to that public reasoning.
180

Conclusion: Evaluation and Valuation

This dissertation has presented and defended a particular interpretation

of Amartya Sen’s capability approach, as a “framework of thought”
rather than a theory of welfare or justice. I will conclude by reflecting on
the theme of evaluation and valuation.
In chapters 2 and 3 I argued firstly against certain ways of
interpreting Sen’s writings on the capability approach, and secondly for
seeing them in the context of Sen’s methodology of evaluation. The
capability approach provides an organising conceptual framework that
can be particularly helpful in trying to understand and make
judgements of individual advantage and social welfare. (The humbleness
of this ambition is somewhat offset by Sen’s claim that capability is a
fundamentally better metric than any of its competitors.) Sen notes that
his account is a general approach “concerned with showing the cogency
of a particular space for the evaluation of individual opportunities and
successes” (Sen 1993b, 49 original emphases) compatible with various
routes to operationalization (such as Nussbaum’s theory of justice). I
argued that the “deliberate incompleteness” of his account reflects Sen’s
general concern and preoccupation with the demands of evaluation.
Evaluations require making choices about how best to understand a
particular case through reflecting on which information and theoretical
perspectives are salient. And these choices are the responsibility of the
individual doing the evaluation, not the theories concerned. They
require the exercise of judgement rather than the application of a
formula, and thus an active, expert, honest evaluator or spectator.
Evaluation can be distinguished from valuation. While normative
evaluation is concerned with finding out the degree to which certain
normative standards or values are achieved, valuation is concerned with
the logically prior exercise of determining those standards and values.
Sen’s arguments for leaving valuation to individual reflection and the
exercise of social choice seem intended to link these distinct operations
to distinct kind of persons (i.e. ‘actors’ as opposed to ‘spectators’). This
relates to Sen’s concern with what might be called a social science ethic,
not always shared by moral philosophers, that evaluators should not
engage in the exercise of “unleashing one’s morals on the statistics” (Sen
 181

1980, 366).199 It also relates to Sen’s concern that technocrats should not
attempt to derive simple answers to complex valuational questions
through an inappropriate use of such limited statistical methods as
contingent valuation or real income metrics (Sen 1999a, 79–81).
This distinction between valuation and evaluation that I have
suggested is implicit in Sen’s writing is itself a rather Sennian one. Both
actors and spectators are engaged in what I analysed as “judgement” –
the exercise of critical reflection about how to treat certain issues or
answer certain questions. Furthermore, spectators are often explicitly
engaged in valuational exercises required by the methodology of
evaluation (such as which capabilities to include in an index like the
Human Development Index, intended to provide a perspective on how
well people’s lives are going, and how to weight them). Actors, too, are
often explicitly concerned with evaluation, such as in the public
reasoning aspect of social choice (which complements the explicit
choosing - ‘voting’ - aspect, in Sen’s account).
The difference between valuation and evaluation is thus founded not
on a dichotomy between two different kinds of reasoning, but rather on
an ethical distinction between the appropriate roles of those involved.
The job of spectators is the scientific assessment of how things are
going, to produce findings that are objective in the sense that their
methodology could withstand critical scrutiny. Doing this job correctly
requires abstaining from inserting oneself and one’s own moral values
into the analysis. That would be unethical in two senses. It would be an
exercise of asserting one’s own subjective and contentious moral views
onto the world without subjecting them to the scrutiny of others (a
legitimacy problem), and it would not be a good evaluation to give
because it would not relate to the concerns of those involved (an
epistemic problem).
Spectators, at least the ones who write government reports and
academic papers, have a privileged position as those with the resources
(such as information and sophisticated theoretical perspectives) to gain
an overview of how well and how badly a society is doing, and also the
opportunity to influence how society will go if their evaluations
influence or justify public policy. But nothing in that privileged role
gives them the right or the ability to decide what people should or

199
For example, “For the person studying and measuring poverty, the conventions of
society are matters of fact (what are the contemporary standards?) and not issues of
morality or of subjective search (what should be the contemporary standards? what
should be [my] values? how do I feel about all this?).” (Sen 1979e, 285–6)
182

shouldn’t value. This issue may be evaded by resolute attachment to

particularly primitive evaluational methods, such as the self-limitations
of orthodox welfare economics or the focus on economic production
statistics in development economics. But it is particularly obvious where
capability is the focal space of evaluation, for there is something
nonsensical in attempting to evaluate someone’s opportunities to live
different kinds of life without considering the value of those
opportunities to them, “without some idea of what a person prefers and
has reason to prefer" (Sen 2002e, 5). Thus, pace critics such as Sugden,
Sen argues that the evaluative exercise of the spectator must be
subordinate to the valuational exercise of those concerned, considered
as “agents” rather than “patients”. In terms of the central ethical
question “who decides?”, the actors decide what kind of life is a good
one, while the spectators decide how best to evaluate how well people
are doing in those terms. The priority of the actors is further asserted in
their oversight of the methodology of evaluation, which Sen argues,
perhaps overoptimistically, should be made explicit so that it can be the
object of public reason and scrutiny (e.g. Sen 1999a, 81).
The second half of this dissertation was particularly concerned with
valuation issues and the complicated roles of actors and spectators. In
chapter 4, I focused on the problem of adaptation, which I argued is
understood by Sen in terms of the deprivation of practical reasoning
(“the fourth aspect of the self”) in reaction to material deprivation and
social oppression. This poses a challenge to the distinction between
evaluation and valuation, spectators and actors, because, if actors’ self-
reasoning is distorted in ways that reflect the domination of
circumstances, then their subjective perspectives of ‘what kind of life
they have reason to value’ become unreliable. Yet, the spectators would
be overstepping their proper role if they attempted to decide on that
question on behalf of those concerned. There are various ways the
capability approach can still be applied. For example, it can still
diagnose and problematise achievements and failings in the dimension
of well-being (such as the relative undernourishment of girls with
respect to boys). It can also examine the determinants of opportunities
for living certain kinds of lives (such as social or legal restrictions on the
schooling or employment of members of certain demographic groups)
and thus analyse the limits that exist on people’s freedom of choice,
whatever they value. But it cannot say what people should value.
 183

I argued that Adam Smith’s device of the impartial spectator could

be of some use in addressing this challenge of adaptation. First, by
conceptualising the problem in terms of values, desires, and aspirations
that would not survive transpositional scrutiny (in other words, which
the impartial spectator would not be able to go along with). Second, by
directing the spectator to adopt the ethos of the impartial spectator and
attend carefully and respectfully to the close scrutiny of the
‘reasonableness’ of the relation between people’s preferences and their
circumstances. Though in many cases the spectator will not be able to
decide whether or not adaptation has occurred, in at least some cases
such an evaluation may reasonably lead to clear judgements one way or
another.
Besides improving spectatorship, this conceptualisation of
adaptation also points to how the position of actors may be improved,
by supporting or enhancing their capability to critically evaluate and
judge their own preferences. Sen argues that “the ability to doubt and to
question is within each person’s capacity” and that this is one of the
things that distinguish us from other animals, that make us human (Sen
1999d, 24). Yet translating this innate capacity for critical reasoning into
a capability requires opportunities for practise. Sen therefore argues for
“the creation of conditions in which people have real opportunities of
judging the kind of lives they would like to lead” (Sen 1999a, 63). The
impartial spectator framework points to an important kind of
opportunity, access to multiple alternative epistemic positions - in
particular, other lives - from which one’s own values can be considered
and scrutinised in transpositional terms. Providing people with
opportunities to take up the perspectives of other people on their lives
or their values would seem to be an important tool for remediating or
preventing adaptation while respecting the fundamental autonomy of
the actors concerned.
Chapter 5 shifted the focus from the evaluation of advantage to
more conventional issues in development ethics, regarding how
transformative interventions in people’s lives can be justified and the
constraints those justifications imply. Again, the distinction between
valuation and evaluation is crucial. The spectator who evaluates
individual advantage and social welfare in terms of capability must
consider not only the comparison of states before and after an
intervention, but also the process by which the latter is brought about.
Prioritising the valuational perspective of the actors in development
184

contexts requires that development interventions be justified to them in

advance.
Because the evaluation of people’s capability relates to the kind of
life they have reason to value, the capability approach is fundamentally
“agent-oriented”. Because people are expected to change their notions
about the kind of life they have reason to value over the course of
development, development is transformational: it changes the standards
by which its success or failure can be judged. Such shifting standards
make it problematic to evaluate and justify human development
interventions in terms of before-and-after-comparisons of capability
sets.
In particular, one cannot claim that just because the people who
took part in a project genuinely valued its outcomes in terms of
enhancing their capability, that one has achieved human development.
Without consideration of how the project was justified to the people in
the first place one cannot tell whether it was genuinely agent-oriented,
respecting the priority of the actors in determining what values count. If
a development project can be justified by evaluations of states of affairs
which haven’t happened yet, this gives those who conduct the
evaluations a great deal of liberty to determine the values they are based
on. They can claim that development projects will bring about certain
kinds of people who will value what they have become (such as literate,
egalitarian, religious, and so forth). Such an attitude is a “patient”
oriented one that sees people “primarily as passive recipients of the
benefits of cunning development programs” (Sen 1999a, 11). Genuinely
human development, I argue, must instead engage with individuals as
autonomous agents in managing their own transformation. This
requires that the justifications offered for development be ones that the
people concerned (rather than future versions of them) have reason to
value.
As I noted above, the two kinds of person associated with the
operations of evaluation and valuation, ‘spectators’ and ‘actors’, seem
modelled, respectively, on the empirical social scientist and the ordinary
persons they study (exercising their own capability for practical reason).
It would seem that there is not much of a role left for the moral
philosopher in this scheme (beyond Sen’s proposal of the intrinsic
importance of capability and reflections on methodology). This is
apparent in several ways. The social-scientific orientation of Sen’s
concern (which I discussed in chapters 2 and 3) means that beyond a
 185

few rather basic, and perhaps even commonsensical, claims about the
moral significance of what we can actually be and do, it is the
methodology of evaluation that is the main focus of discussion. It is
therefore difficult for moral philosophers (such as Nussbaum) to insert
the analysis of value claims into the debate. The “agent-oriented”
character of the approach also limits the role of philosophical theory,
since it asserts the priority of the actors’ valuational perspective, singly
or collectively, over the virtues of moral theory (such as truth).
Philosophers are free to make suggestions and criticisms to the actors
about what kinds of life really are valuable (in the role of what Claassen
calls “philosopher citizens”), but their theoretical accounts have no
special status in the decision making process itself.
This exclusion of philosophical theory is perhaps most apparent in
Sen’s approach to thinking of social justice, which I discussed in chapter
6. Sen rejects the weighty philosophical apparatus of social contract
theory in favour of the comparatively slight formal framework of social
choice theory, which he notes is “quite close to the commonsense
understanding of the nature of appropriate social decisions” (Sen 2009a,
18). He also rejects the intellectual satisfactions of identifying what a
truly just society would look like, in favour of a rather more mundane,
and less radical, focus on identifying what are generally agreed to be
injustices. Even for this task, Sen proposes that actors do the work
themselves, through democracy.
In the context of much of the dominant (Anglo-Saxon) tradition of
“political moralism”, this makes Sen a radical democrat. He rejects the
claim that the philosophical spectator’s evaluation of the requirements
of justice (whether substantive theories of the good or theories of the
just use of power) should trump actual democratic practise. Theory has
a descriptive rather than prescriptive role in this analysis of democracy.
Sen makes use of theoretical resources in analysing the deficiencies of
actual democracies (such as India’s) and how they might be improved.
But he argues against the idea that the legitimacy of the outcomes of
democratic practise, or of a political regime itself, depend on the
fulfilment of particular theoretical qualities. Rather, he argues for
appreciating, and promoting, the scope of actual democratic freedoms
and opportunities to support public reasoning and the effective
politicisation of injustices such as gross deprivations.
This thematic review may seem to have cast Sen as an opponent of
philosophy. But I do not believe this impression is correct. Sen’s work
186

can certainly be seen to be influenced by his identity as an accomplished

social scientist, for example, in his concern with methodology, logistical
issues, wily agents, and his disconcertingly pragmatic use of
terminology. But it is equally influenced by a philosopher’s concern for
foundational analysis, nuance, and ethical breadth. My engagement with
Sen’s work has persuaded me that it succeeds because of its deep, and
deeply informed, inter-disciplinarity rather than in spite of it. Sen is a
worldly philosopher, concerned with understanding not only how the
world really is, but how it should be made better.
 187

Bibliography
Alkire, Sabina. 2005. Valuing Freedoms. Oxford University Press.
Alkire, Sabina, and Maria Emma Santos. 2010a. “Acute Multidimensional
Poverty: A New Index for Developing Countries.” Oxford Poverty &
Human Development Initiative (OPHI) Working Paper (38) (July).
http://www.ophi.org.uk/acute-multidimensional-poverty-a-new-index-
for-developing-countries/.
———. 2010b. “Research Brief: Multidimensional Poverty Index”. Oxford
Poverty and Human Development Initiative (OPHI).
http://www.ophi.org.uk/wp-content/uploads/OPHI-MPI-Brief.pdf.
Anderson, Elizabeth. 1999. “What Is the Point of Equality?” Ethics 109 (2): 287–
337.
———. 2001. “Symposium on Amartya Sen’s Philosophy: 2 Unstrapping the
Straitjacket of ‘preference’: a Comment on Amartya Sen’s Contributions
to Philosophy and Economics.” Economics and Philosophy 17 (01): 21–38.
———. 2003. “Sen, Ethics, and Democracy.” Feminist Economics 9 (2): 239–261.
———. 2010. “Justifying the Capability Approach to Justice.” In Measuring
Justice, ed. Harry Brighouse and Ingrid Robeyns, 81–100. Cambridge
University Press.
Apffel-Marglin, Frédérique. 2010. “Questioning the Capability Approach.” In
Interrogating Development: Insights from the Margins, ed. Frédérique
Apffel-Marglin, Sanjay Kumar, and Arvind Mishra, 209–216. Oxford
University Press.
Arneson, Richard J. 2010. “Two Cheers for Capabilities.” In Measuring Justice,
ed. Harry Brighouse and Ingrid Robeyns, 101–128. Cambridge University
Press.
Arrow, Kenneth J. 1951. Social Choice and Individual Values. New York: Wiley.
———. 1999. “Amartya K. Sen’s Contributions to the Study of Social Welfare.”
Scandinavian Journal of Economics 101 (2): 163-172.
Atkinson, Anthony B. 1999. “The Contributions of Amartya Sen to Welfare
Economics.” The Scandinavian Journal of Economics 101 (2): 173–190.
Banerjee, Abhijit V., and Esther Duflo. 2011. Poor Economics: A Radical
Rethinking of the Way to Fight Global Poverty. Public Affairs.
Basu, Kaushik. 2011. “Why, for a Class of Bribes, the Act of Giving a Bribe
Should Be Treated as Legal.” India Ministry of Finance Working Paper.
http://finmin.nic.in/workingpaper/act_giving_bribe_legal.pdf.
Becker, Gary Stanley. 1996. Accounting for Tastes. Harvard University Press.
Bell, Daniel. 2012. “Communitarianism.” In The Stanford Encyclopedia of
Philosophy, ed. Edward N. Zalta. Spring 2012.
http://plato.stanford.edu/archives/spr2012/entries/communitarianism
/.
Bhatia, Bela, and Jean Drèze. 2006. “Employment Guarantee on the Ground:
Insights from Jharkhand.” Economic and Political Weekly (July 22): 3198–
3202.
Blomfield, Megan. 2012. “Ethics in Economics: Lessons from Human Subjects
Research.” Erasmus Journal for Philosophy and Economics 5 (1): 24–44.
Bohman, James. 1997. “Deliberative Democracy and Effective Social Freedom:
Capabilities, Resources, and Opportunities.” In Deliberative Democracy:
Essays on Reason and Politics, ed. James Bohman and William Rehg, 321–
348. MIT Press.
188

Boumans, Marcel, and Mary S. Morgan. 2001. “Ceteris Paribus Conditions:

Materiality and the Application of Economic Theories.” Journal of
Economic Methodology 8 (1): 11–26.
Broadie, Alexander. 2006. “Sympathy and the Impartial Spectator.” In The
Cambridge Companion to Adam Smith, ed. Knud Haakonssen, 158–188.
Cambridge University Press.
Buchanan, James M. 1954. “Social Choice, Democracy, and Free Markets.” The
Journal of Political Economy 62 (2): 114–123.
Claassen, Rutger. 2011. “Making Capability Lists: Philosophy Versus
Democracy.” Political Studies 59 (3): 491–508.
Claassen, Rutger, and Marcus Düwell. 2012. “The Foundations of Capability
Theory: Comparing Nussbaum and Gewirth.” Ethical Theory and Moral
Practice (April 11).
Clark, David A. 2002. Visions of Development: a Study of Human Values. Edward
Elgar Publishing.
———. 2009. “Adaptation, Poverty and Well‐Being: Some Issues and
Observations with Special Reference to the Capability Approach and
Development Studies.” Journal of Human Development and Capabilities
10 (1): 21–42.
———. 2012. “Adaptation and Development - Issues, Evidence and Policy
Relevance.” In Adaptation, Poverty and Development: The Dynamics of
Subjective Well-Being, 1–31. Palgrave Macmillan.
Clark, David A., and Mozaffar Qizilbash. 2008. “Core Poverty, Vagueness and
Adaptation: A New Methodology and Some Results for South Africa.”
Journal of Development Studies 44 (4): 519–544.
Cohen, G.A. 1993. “Equality of What? On Welfare, Goods, and Capabilities.” In
The Quality of Life, ed. Martha Nussbaum and Amartya Sen, 9–29.
Oxford University Press.
———. 1994. “Amartya Sen’s Unequal World.” New Left Review.
Conradie, Ina, and Ingrid Robeyns. Forthcoming. “Aspirations and Human
Development Interventions.” Journal of Human Development and
Capabilities.
Crick, Bernard. 2002. Democracy: A Very Short Introduction. Oxford University
Press.
Crocker, David A. 2009. Ethics of Global Development: Agency, Capability, and
Deliberative Democracy. Cambridge University Press.
Daniels, N. 2006. “Toward Ethical Review of Health System Transformations.”
Journal Information 96 (3).
Dasgupta, Partha. 2005. “What Do Economists Analyze and Why: Values or
Facts?” Economics and Philosophy 21 (02): 221–278.
———. 2009. “Facts and Values in Modern Economics.” In The Oxford
Handbook of Philosophy of Economics, ed. Harold Kincaid and Don Ross,
580–640. Oxford University Press US.
Davis, John B. 2012. “The Idea of Public Reasoning.” Journal of Economic
Methodology 19 (2): 169–172.
Deaton, Angus, and Jean Dreze. 2009. “Food and Nutrition in India: Facts and
Interpretations.” Economic and Political Weekly 44 (7): 42–65.
DeMartino, George. 2011. The Economist’s Oath: On the Need for and Content of
Professional Economic Ethics. Oxford University Press.
 189

Deneulin, Séverine. 2002. “Perfectionism, Paternalism and Liberalism in Sen

and Nussbaum’s Capability Approach.” Review of Political Economy 14
(4): 497.
———. 2008. “Beyond Individual Freedom and Agency: Structures of Living
Together in the Capability Approach.” In The Capability Approach, ed.
Flavio Comim, Mozaffar Qizilbash, and Sabina Alkire, 105–124.
Cambridge University Press.
Deveaux, Monique. 2002. “Political Morality and Culture: What Difference Do
Differences Make?” Social Theory and Practice 28 (3): 503–518.
Dewey, John. 1954. The Public and Its Problems.
Dey, N., Jean Dreze, and R. Khera. 2006. Employment Guarantee Act: a Primer.
New Delhi: National Book Trust.
Dowding, Keith. 2006. “Can Capabilities Reconcile Freedom and Equality?”
Journal of Political Philosophy 14 (3): 323–336.
Drèze, Jean, and Amartya Sen. 1989. Hunger and Public Action. Oxford
University Press.
———. 2002. India: Development and Participation. Oxford University Press.
Dryzek, John S., and Christian List. 2003. “Social Choice Theory and
Deliberative Democracy: A Reconciliation.” British Journal of Political
Science 33 (1): 1–28.
Duflo, Esther. 2012. “Human Values and the Design of the Fight Against
Poverty”. Tanner Lectures. Harvard University.
http://www.povertyactionlab.org/sites/default/files/documents/Tanner
Lectures_EstherDuflo_draft.pdf.
Dworkin, Ronald. 2002. Sovereign Virtue: The Theory and Practice of Equality.
Harvard University Press.
Elster, Jon. 1982. “Sour Grapes - Utilitarianism and the Genesis of Wants.” In
Utilitarianism and Beyond, ed. Amartya Sen and Bernard Williams, 219–
238. Cambridge University Press.
Eyal, Nir. 2011. “Informed Consent.” In The Stanford Encyclopedia of
Philosophy, ed. Edward N. Zalta. Fall 2012.
http://plato.stanford.edu/archives/fall2012/entries/informed-consent/.
Farrelly, Colin. 2007. “Justice in Ideal Theory: A Refutation.” Political Studies 55
(4): 844–864.
Fishkin, James S. 1991. Democracy and Deliberation. Yale University Press.
Fleischacker, Samuel. 1999. A Third Concept of Liberty. Princeton University
Press.
Foster, James, and Christopher Handy. 2009. “External Capabilities.” In
Arguments for a Better World: Essays in Honor of Amartya Sen, ed.
Kaushik Basu and Ravi Kanbur, 1:363–374. Oxford University Press US.
Gasper, Des. 1993. “Entitlements Analysis: Relating Concepts and Contexts.”
Development and Change 24 (4): 679–718.
———. 1997. “Sen’s Capability Approach and Nussbaum’s Capabilities Ethic.”
Journal of International Development 9 (2): 281–302.
———. 2002. “Is Sen’s Capability Approach an Adequate Basis for Considering
Human Development?” Review of Political Economy 14 (4): 435–461.
Gaus, Gerald. 2012. “Social Contract and Social Choice.” Rutgers Law Review 43
(2): 243–276.
Gibbard, Allan. 1973. “Manipulation of Voting Schemes: A General Result.”
Econometrica 41 (4): 587.
190

Giri, Ananta Kumar. 2000. “Rethinking Human Well-being: a Dialogue with

Amartya Sen.” Journal of International Development 12 (7): 1003–1018.
Goodland, Robert. 2004. “Free, Prior and Informed Consent and the World
Bank Group.” Sustainable Development Law & Policy 4: 66-74.
Gore, Charles. 1997. “Irreducibly Social Goods and the Informational Basis of
Amartya Sen’s Capability Approach.” Journal of International
Development 9 (2): 235–250.
Goulet, Denis. 1992. “Development: Creator and Destroyer of Values.” World
Development 20 (3): 467–475.
Green, Duncan. 2008. From Poverty to Power: How Active Citizens and Effective
States Can Change the World. Oxfam.
Griswold, Charles L. 1999. Adam Smith and the Virtues of Enlightenment.
Cambridge University Press.
Gutmann, Amy, and Dennis Frank Thompson. 1996. Democracy and
Disagreement. Harvard University Press.
Herdt, Tom De, and Séverine Deneulin. 2007. “Guest Editors’ Introduction.”
Journal of Human Development 8 (2): 179–184.
Hirschman, Albert O. 1992. “A Dissenter’s Confession: The Strategy of
Economic Development Revisited.” In Rival Views of Market Society and
Other Recent Essays, 3–34. Harvard University Press.
Honig, Bonnie. 1993. Political Theory and the Displacement of Politics. Cornell
University Press.
Jerven, Morten. 2013. Poor Numbers: How We Are Misled by African
Development Statistics and What to Do About It. Cornell University Press.
Kaufman, Alexander. 2006. “Capabilities and Freedom.” Journal of Political
Philosophy 14 (3): 289–300.
Keirns, Pamela. 2008. Water Supply and Sanitation Services for the Rural Poor:
The Gram Vikas Experience. Practical Action Publishing.
Kelly, Erin. 2012. “Public Reason as a Collective Capability.” Rutgers Law
Review 43 (2): 295–316.
Klugman, J., F. Rodríguez, and H. J Choi. 2011. “The HDI 2010: New
Controversies, Old Critiques.” Human Development Research Papers
(2009 to Present).
Knight, Jack, and James Johnson. 1997. “What Sort of Equality Does
Deliberative Democracy Require?” In Deliberative Democracy: Essays on
Reason and Politics, ed. James Bohman and William Rehg, 279–321. MIT
Press.
Krause, Sharon R. 2013. “Beyond Non-domination Agency, Inequality and the
Meaning of Freedom.” Philosophy & Social Criticism 39 (2): 187–208.
Kuklys, Wiebke. 2005. Amartya Sen’s Capability Approach. Springer.
Kynch, J, and Amartya Sen. 1983. “Indian Women: Well-being and Survival.”
Cambridge Journal of Economics 7 (3-4): 363–380.
Lelyveld, Joseph. 2011. Great Soul: Mahatma Gandhi and His Struggle with
India. Knopf Doubleday Publishing Group.
Lin, Justin Yifu, and Dennis Tao Yang. 2000. “Food Availability, Entitlements
and the Chinese Famine of 1959-61.” The Economic Journal 110 (460):
136–158.
Lipsey, Richard G. 2007. “Reflections on the General Theory of Second Best at
Its Golden Jubilee.” International Tax and Public Finance 14 (4): 349–
364.
 191

Lipsey, Richard G., and Kelvin Lancaster. 1956. “The General Theory of Second
Best.” The Review of Economic Studies 24 (1): 11–32.
List, Christian, Robert C. Luskin, James S. Fishkin, and Iain McLean. 2006.
“Deliberation, Single-Peakedness, and the Possibility of Meaningful
Democracy: Evidence From Deliberative Polls.”
Livet, Pierre. 2006. “Identities, Capabilities and Revisions.” Journal of Economic
Methodology 13 (3): 327–348.
MacAuslan, Ian. 2007. India’s National Rural Employment Guarantee Act: A
Case Study for How Change Happens. Oxfam International.
MacIntyre, Alasdair C. 1981. After Virtue. Duckworth.
Mackie, Penelope. 2005. The Oxford Companion to Philosophy. Ed. Ted
Honderich. 2nd ed. Oxford University Press.
Marglin, Stephen A. 1990. “Towards the Decolonization of the Mind.” In
Dominating Knowledge: Development, Culture, and Resistance, ed.
Frédérique Apffel Marglin and Stephen A. Marglin, 1–28. Clarendon
Press.
———. 2010. “From Imperialism to Globalization, by Way of Development.” In
Interrogating Development: Insights from the Margins, ed. Frédérique
Apffel-Marglin, Sanjay Kumar, and Arvind Mishra, 287–307. Oxford
University Press.
Menon, Nivedita. 2002. “Universalism Without Foundations?” Economy and
Society 31 (1): 152-169.
Mill, John Stuart. 1859 [1998]. “On Liberty.” In On Liberty and Other Essays, ed.
John Gray. Oxford World Classics
———. 1861 [1998]. “Considerations of Representative Government.” In On
Liberty and Other Essays, ed. John Gray. Oxford World Classics
———. 1869 [1998]. “The Subjection of Women.” In On Liberty and Other
Essays, ed. John Gray. Oxford World Classics
Mills, Charles W. 2005. “‘Ideal Theory’ as Ideology.” Hypatia 20 (3): 165–183.
Nafziger, E. Wayne. 2006. Economic Development. 4th ed. Cambridge University
Press.
Nozick, Robert. 1974. Anarchy, State, and Utopia. Basic Books.
Nussbaum, Martha. 1988. “Nature, Function, and Capability: Aristotle on
Political Distribution.” In Oxford Studies in Ancient Philosophy. Oxford
University Press.
———. 1993. “Non-relative Virtues: An Aristotelian Approach.” In The Quality
of Life, ed. Amartya Sen and Martha Nussbaum. Oxford University Press.
———. 1995. Poetic Justice. Beacon Press.
———. 1997. Cultivating Humanity: A Classical Defense of Reform in Liberal
Education. Harvard University Press.
———. 2000. Women and Human Development. Cambridge University Press.
———. 2001. “Symposium on Amartya Sen’s Philosophy: 5 Adaptive
Preferences and Women’s Options.” Economics and Philosophy 17 (01):
67–88.
———. 2003. “Capabilities as Fundamental Entitlements: Sen and Social
Justice.” Feminist Economics 9 (2): 33-59.
———. 2004. “On Hearing Women’s Voices: A Reply to Susan Okin.” Philosophy
and Public Affairs 32 (2): 193–205.
———. 2005. “Women’s Bodies: Violence, Security, Capabilities.” Journal of
Human Development 6 (2) (July): 167–183.
192

———. 2009a. “Capabilities and Constitutional Law: ‘Perception’ Against Lofty

Formalism.” Journal of Human Development and Capabilities 10 (3): 341.
———. 2009b. “The Capability of Philosophy: An Interview with Martha C.
Nussbaum.” The Minnesota Review, November 14.
———. 2011a. Creating Capabilities: The Human Development Approach.
Harvard University Press.
———. 2011b. “Perfectionist Liberalism and Political Liberalism.” Philosophy &
Public Affairs 39 (1): 3–45.
Nussbaum, Martha, and Amartya Sen. 1989. “Internal Criticism and Indian
Rationalist Traditions.” In Relativism: Interpretation and Confrontation,
ed. Michael Krausz, 299–325. University of Notre Dame Press.
Okin, Susan Moller. 2003. “Poverty, Well-Being, and Gender: What Counts,
Who’s Heard?” Philosophy and Public Affairs 31 (3) (July): 280–316.
Oosterlaken, Ilse. 2012. “Is Pogge a Capability Theorist in Disguise?” Ethical
Theory and Moral Practice (February 22).
Orshansky, Mollie. 1969. “How Poverty Is Measured.” Monthly Labor Review 92
(2): 37.
Osmani, Siddiq. 1995. “The Entitlement Approach to Famine: An Assessment.”
In Choice, Welfare, and Development, ed. Kaushik Basu, Prasanta K.
Pattanaik, and Kōtarō Suzumura, 253–294. Oxford University Press.
———. 2009. “The Sen System of Social Evaluation.” In Arguments for a Better
World: Essays in Honor of Amartya Sen, ed. Kaushik Basu and Ravi
Kanbur, 15–34. Oxford University Press US.
Peter, Fabienne. 2006. “The Political Egalitarian’s Dilemma.” Ethical Theory and
Moral Practice 10 (4): 373–387.
Pettit, Philip. 2001. “Symposium on Amartya Sen’s Philosophy: 1 Capability and
Freedom: a Defence of Sen.” Economics and Philosophy 17 (1): 1–20.
Pitts, Jennifer. 2005. A Turn to Empire: The Rise of Imperial Liberalism in
Britain and France. Princeton University Press.
Pogge, Thomas. 2002. “Can the Capability Approach Be Justified?”
Philosophical Topics 30 (2): 167–228.
———. 2010. “A Critique of the Capability Approach.” In Measuring Justice, ed.
Harry Brighouse and Ingrid Robeyns, 17–60. Cambridge University Press.
Pratichi Trust. 2005a. The Mid-day Meal Programme in West Bengal. New Delhi:
The Pratichi (India) Trust.
———. 2005b. The Pratichi Health Report. New Delhi: The Pratichi (India) Trust.
———. 2009. The Pratichi Education Report II. New Delhi: The Pratichi (India)
Trust.
Putnam, Hilary. 2002. The Collapse of the Fact/value Dichotomy and Other
Essays. Harvard University Press.
———. 2008. “Capabilities and Two Ethical Theories.” Journal of Human
Development 9 (3): 377–388.
Qizilbash, Mozaffar. 2005. “Sen on Freedom and Gender Justice.” Feminist
Economics 11 (3): 151–166.
———. 2006. “Well-being, Adaptation and Human Limitations.” Royal Institute
of Philosophy Supplements 81: 83–110.
———. 2007. “Social Choice and Individual Capabilities.” Politics, Philosophy &
Economics 6 (2) (June 1): 169–192.
———. 2008. “Amartya Sen’s Capability View: Insightful Sketch or Distorted
Picture?” In The Capability Approach: Concepts, Measures and
 193

Applications, ed. Flavio Comim, Mozaffar Qizilbash, and Sabina Alkire,

53–81. Cambridge University Press.
———. 2009. “The Adaptation Problem, Evolution and Normative Economics.”
In Arguments for a Better World: Essays in Honor of Amartya Sen, ed.
Kaushik Basu and Ravi Kanbur, 50–67. Oxford University Press US.
Rana, Kumar, and Achin Chakraborty, ed. 2009. “The Pratichi Child Report: A
Study on the Delivery of ICDS in West Bengal”.
Rana, Kumar, Abdur Rafique, and Amrita Sengupta. 2002. The Pratichi
Education Report I: The Delivery of Primary Education: a Study in West
Bengal. TLM Books in with association with Pratichi (India) Trust.
Raphael, David Daiches. 2007. The Impartial Spectator: Adam Smith’s Moral
Philosophy. Oxford University Press.
Rawls, John. 1985. “Justice as Fairness: Political Not Metaphysical.” Philosophy
and Public Affairs 14 (3): 223–251.
———. 1999. A Theory of Justice (TJ): Revised Edition. Harvard University Press.
———. 2003a. “The Law of Peoples.” In The Law of Peoples with “The Idea of
Public Reason Revisited”. Harvard University Press.
———. 2003b. “The Idea of Public Reason Revisited.” In The Law of Peoples
with “The Idea of Public Reason Revisited”. Harvard University Press.
———. 2005. Political Liberalism. Expanded ed. New York: Columbia University
Press.
Roberts, C. J. 2007. Parents Involved in Community Schools V. Seattle School
Dist. No. 1 (Chief Justice Roberts, Opinion of the Court) U.S. U.S. Supreme
Court.
Robeyns, Ingrid. 2000. “An Unworkable Idea or a Promising Alternative?: Sen’s
Capability Approach Re-examined.” Katholieke Universiteit Leuven:
Center for Economic Studies Discussions Paper Series (DPS) (00.30).
———. 2003. “Sen’s Capability Approach and Gender Inequality: Selecting
Relevant Capabilities.” Feminist Economics 9 (2): 61–92.
———. 2005. “Selecting Capabilities for Quality of Life Measurement.” Social
Indicators Research 74 (1): 191–215.
———. 2006. “The Capability Approach in Practice.” Journal of Political
Philosophy 14 (3): 351–376.
———. 2008. “Ideal Theory in Theory and Practice.” Social Theory and Practice
34 (3): 341–362.
———. 2009. “Justice as Fairness and the Capability Approach.” In Arguments
for a Better World: Essays in Honor of Amartya Sen, ed. Kaushik Basu
and Ravi Kanbur, 1:397–413. Oxford University Press US.
———. 2012. “Are Transcendental Theories of Justice Redundant?” Journal of
Economic Methodology 19 (2): 159–163.
Roemer, John E. 1998. Theories of Distributive Justice. Harvard University Press.
Roy, Arundhati. 1999. The Cost of Living. Modern Library.
Sandel, Michael. 2012. What Money Can’t Buy: The Moral Limits of Markets.
Penguin Books Limited.
Satterthwaite, Mark Allen. 1975. “Strategy-proofness and Arrow’s Conditions:
Existence and Correspondence Theorems for Voting Procedures and
Social Welfare Functions.” Journal of Economic Theory 10 (2): 187–217.
Satz, Debra. 2012. “Amartya Sen’s ‘The Idea of Justice’: What Approach, Which
Capabilities?” Rutgers Law Review 43 (2): 277–294.
Scanlon, T. M. 1975. “Preference and Urgency.” The Journal of Philosophy 72
(19): 655–669.
194

Schumacher, Reinhard. 2012. “Adam Smith’s Theory of Absolute Advantage

and the Use of Doxography in the History of Economics.” Erasmus
Journal for Philosophy and Economics 5 (2): 54–80.
Sen, Amartya. 1966. “Hume’s Law and Hare’s Rule.” Philosophy 41 (155): 75–79.
———. 1967. “The Nature and Classes of Prescriptive Judgements.” The
Philosophical Quarterly 17 (66): 46–62.
———. 1970a. “The Impossibility of a Paretian Liberal.” The Journal of Political
Economy 78 (1): 152–157.
———. 1970b. Collective Choice and Social Welfare. Holden-Day.
———. 1977. “Rational Fools: A Critique of the Behavioral Foundations of
Economic Theory.” Philosophy and Public Affairs 6 (4): 317–344.
———. 1979a. “Utilitarianism and Welfarism.” The Journal of Philosophy 76 (9):
463–489.
———. 1979b. “Personal Utilities and Public Judgements: Or What’s Wrong
With Welfare Economics.” The Economic Journal 89 (355): 537–558.
———. 1979c. Equality of What? Stanford university: Tanner Lectures on
Human Values.
———. 1979d. “Informational Analysis of Moral Principles.” In Rational Action:
Studies in Philosophy and Social Science, ed. Ross Harrison, 115–132.
Cambridge University Press.
———. 1979e. “Issues in the Measurement of Poverty.” The Scandinavian
Journal of Economics 81 (2): 285–307.
———. 1980. “Description as Choice.” Oxford Economic Papers 32 (3). New
Series: 353–369.
———. 1981. “Ingredients of Famine Analysis: Availability and Entitlements.”
The Quarterly Journal of Economics 96 (3): 433–464.
———. 1982. Poverty and Famines. Oxford University Press.
———. 1983a. “Accounts, Actions and Values: Objectivity of Social Science.” In
Social Theory and Political Practice, ed. C. Lloyd. Oxford: Clarendon
Press.
———. 1983b. “Evaluator Relativity and Consequential Evaluation.” Philosophy
and Public Affairs 12 (2): 113–132.
———. 1983c. “Poor, Relatively Speaking.” Oxford Economic Papers 35 (2): 153–
169.
———. 1984a. “Economics and the Family.” In Resources, Values, and
Development, 369–388. Harvard University Press.
———. 1984b. “Goods and People.” In Resources, Values, and Development,
509–532. Harvard University Press.
———. 1985a. “Well-Being, Agency and Freedom: The Dewey Lectures 1984.”
The Journal of Philosophy 82 (4): 169–221.
———. 1985b. Commodities and Capabilities. North-Holland.
———. 1985c. “The Standard of Living.” Tanner Lectures on Human Values,
Cambridge UK.
———. 1989a. “Development as Capability Expansion.” Journal of Development
Planning 19: 41–58.
———. 1989b. “Reply.” In The Standard of Living, ed. Geoffrey Hawthorn, 103–
112. Cambridge University Press.
———. 1989c. “Cooperation, Inequality, and the Family.” Population and
Development Review 15 (January 1): 61–76.
———. 1990a. “Justice: Means Versus Freedoms.” Philosophy and Public Affairs
19 (2): 111–121.
 195

———. 1990b. “Gender and Cooperative Conflicts.” In Persistent Inequalities:

Women and World Development, ed. Irene Tinker. Oxford University
Press.
———. 1990c. “Individual Freedom as Social Commitment.” India International
Centre Quarterly 17 (1): 101–115.
———. 1990d. “More Than 100 Million Women Are Missing.” The New York
Review of Books, December 20.
———. 1992. Inequality Reexamined. New York  ;Oxford [England]: Russell Sage
Foundation ;;Clarendon Press.
———. 1993a. “Positional Objectivity.” Philosophy and Public Affairs 22 (2):
126–145.
———. 1993b. “Capability and Well-being.” In The Quality of Life, ed. Martha
Nussbaum and Amartya Sen, 30–53. Oxford University Press.
———. 1993c. “Internal Consistency of Choice.” Econometrica 61: 495–495.
———. 1995a. “Environmental Evaluation and Social Choice: Contingent
Valuation and the Market Analogy.” Japanese Economic Review 46 (1):
23–37.
———. 1995b. “Rationality and Social Choice.” American Economic Review 85
(1): 1–24.
———. 1997a. “Maximization and the Act of Choice.” Econometrica 65 (4): 745–
779.
———. 1997b. “Editorial: Human Capital and Human Capability.” World
Development 25 (12): 1959–1961.
———. 1999a. Development as Freedom. Oxford University Press.
———. 1999b. “1998 Nobel Prize for Economics Lecture: The Possibility of
Social Choice.” The American Economic Review 89 (3): 349–378.
———. 1999c. “Democracy as a Universal Value.” Journal of Democracy 10 (3):
3–17.
———. 1999d. Reason Before Identity: The Romanes Lecture for 1998. Oxford
University Press.
———. 2000a. “Consequential Evaluation and Practical Reason.” The Journal of
Philosophy 97 (9): 477–502.
———. 2000b. “The Discipline of Cost‐Benefit Analysis.” The Journal of Legal
Studies 29 (June): 931–952.
———. 2001. “Symposium on Amartya Sen’s Philosophy: 4 Reply.” Economics
and Philosophy 17 (1): 51–66.
———. 2002a. “Freedom and Social Choice: The Arrow Lectures.” In Rationality
and Freedom, 581–712. Cambridge Mass.: Belknap Press.
———. 2002b. “Processes, Liberty and Rights.” In Rationality and Freedom,
623–658. Cambridge Mass.: Belknap Press.
———. 2002c. “Why Health Equity?” Health Economics 11 (8): 659–666.
———. 2002d. “Open and Closed Impartiality.” The Journal of Philosophy 99
(9): 445–469.
———. 2002e. Rationality and Freedom. Cambridge Mass.: Belknap Press.
———. 2002f. “Response to Commentaries.” Studies in Comparative
International Development (SCID) 37 (2): 78–86.
———. 2003a. “Continuing the Conversation: Interview with Amartya Sen.”
Feminist Economics 9 (2): 319.
———. 2003b. “Democracy And Its Global Roots.” The New Republic, October 6.
———. 2004a. “Capabilities, Lists, and Public Reason: Continuing the
Conversation.” Feminist Economics 10 (3): 77–80.
196

———. 2004b. “Economic Methodology: Heterogeneity and Relevance.” Social

Research 71 (3): 583.
———. 2004c. UN Human Development Report 2004: Chapter 1 Cultural Liberty
and Human Development. UN Human Development Reports. United
Nations Development Programme.
———. 2004d. “Elements of a Theory of Human Rights.” Philosophy and Public
Affairs 32 (4): 315–356.
———. 2005. “Why Exactly Is Commitment Important for Rationality?”
Economics and Philosophy 21 (01): 5–14.
———. 2006a. “Why Health Equity?” In Public Health, Ethics, and Equity, ed.
Sudhir Anand, Fabienne Peter, and Amartya Sen, 21–36. Oxford
University Press.
———. 2006b. “What Do We Want from a Theory of Justice?” Journal of
Philosophy 103 (5): 215–238.
———. 2006c. “The Man Without a Plan.” Foreign Affairs (March/April).
———. 2006d. “Reason, Freedom and Well-Being.” Utilitas 18 (01): 80–96.
———. 2006e. “Normative Evaluation and Legal Analogues.” In Norms and the
Law, ed. John N. Drobak, 247–266. Cambridge University Press.
———. 2007. On Ethics and Economics. Reprint. Oxford: Blackwell.
———. 2008a. “The Idea of Justice.” Journal of Human Development 9 (3): 331.
———. 2008b. “Hiren Mukherjee Memorial Lecture: The Demands of Social
Justice” August 11, Indian Parliament.
http://cambridgeforecast.wordpress.com/2008/08/26/professor-
amartya-sen-lecture-at-the-indian-parliament/.
———. 2009a. The Idea of Justice. Harvard University Press.
———. 2009b. “Learning from Successes and Failures.” The Hindu, December
19. http://www.thehindu.com/opinion/op-ed/article67082.ece.
———. 2009c. “Primary Schooling: I - A Report on Improvments and Problems
in West Bengal Schools.” The Telegraph, December 19.
http://www.telegraphindia.com/1091219/jsp/opinion/story_11882013.j
sp.
———. 2009d. “Behind the Differential Reach of Primary Education.” The
Hindu, December 20. http://www.thehindu.com/opinion/op-
ed/article67453.ece.
———. 2010a. “Adam Smith and the Contemporary World.” Erasmus Journal
for Philosophy and Economics 3 (1): 50–67.
———. 2010b. “The Place of Capabilities in a Theory of Justice.” In Measuring
Justice, ed. Harry Brighouse and Ingrid Robeyns, 239–253. Cambridge
University Press.
———. 2012a. “Values and Justice.” Journal of Economic Methodology 19 (2):
101–108.
———. 2012b. “A Reply to Robeyns, Peter and Davis.” Journal of Economic
Methodology 19 (2): 173–176.
Sen, Amartya, and Sunil Sengupta. 1983. “Malnutrition of Rural Children and
the Sex Bias.” Economic and Political Weekly 18 (19/21): 855–864.
Sen, Amartya, and Bernard Williams, ed. 1982. Utilitarianism and Beyond.
Cambridge University Press.
Sen, Amartya, and James D. Wolfensohn. 1999. “Let’s Respect Both Sides of the
Development Coin.” The New York Times, May 5.
Shapiro, Ian. 2011. “Review of The Idea of Justice by Amartya Sen.” Journal of
Economic Literature 49 (4): 1251–63.
 197

Sigler, Mary. 1997. “A Sneeze and a Cup of Sugar: A Cautionary Tale of

Narrative and the Law.” Law and Philosophy 16 (6): 617–628.
Singhal, Arvind. 2010. Riding High on Taru Fever: Entertainment-Education
Broadcasts, Ground Mobilization, and Service Delivery in Rural India.
Entertainment-Education and Social Change Wisdom Series. Oxfam
Novib.
Smith, Adam. 1759-1790 [1976]. The Theory of Moral Sentiments (TMS). Ed.
David Daiches Raphael and A. L. Macfie. Indianapolis: Liberty Classics.
———. 1776-1791 [1976]. An Inquiry into the Nature and Causes of the Wealth
of Nations (WN). Ed. Campbell, R. H., Skinner, A. S., and Todd, W. B.
Indianapolis Ind.: Liberty Press.
“Soap Operas - Art Imitating Life.” 2012. BBC World Service Documentary (part
1). http://www.bbc.co.uk/programmes/p00qpj8q.
Srinivasan, Sharath. 2007. “No Democracy Without Justice: Political Freedom in
Amartya Sen’s Capability Approach.” Journal of Human Development 8
(3): 457.
Srinivasan, T. N. 1994. “Human Development: A New Paradigm or Reinvention
of the Wheel?” The American Economic Review 84 (2): 238–243.
St. Clair, Asunción Lera. 2007. “A Methodologically Pragmatist Approach to
Development Ethics.” Journal of Global Ethics 3 (2): 143–164.
Stewart, Frances. 2001. “Review of Women and Human Development: The
Capabilities Approach, by Martha Nussbaum.” Journal of International
Development 13 (8): 1191–1192.
———. 2007. “Basic Needs Approach.” In The Elgar Companion to Development
Studies, ed. David Alexander Clark, 14–18. Edward Elgar Publishing.
Stewart, Frances, and Severine Deneulin. 2002. “Amartya Sen’s Contribution to
Development Thinking.” Studies in Comparative International
Development (SCID) 37 (2): 61–70.
Sugden, Robert. 1993. “Review: Welfare, Resources, and Capabilities: A Review
of Inequality Reexamined by Amartya Sen.” Journal of Economic
Literature 31 (4): 1947–1962.
———. 2006. “What We Desire, What We Have Reason to Desire, Whatever We
Might Desire: Mill and Sen on the Value of Opportunity.” Utilitas 18 (1):
33–51.
Swift, Adam. 2008. “The Value of Philosophy in Nonideal Circumstances.”
Social Theory and Practice 34 (3): 363–387.
Taylor, Charles. 1995. “Irreducibly Social Goods.” In Philosophical Arguments,
127–145. Harvard University Press.
Teschl, Miriam, and Flavio Comim. 2005. “Adaptive Preferences and
Capabilities: Some Preliminary Conceptual Explorations.” Review of
Social Economy 63 (2): 229–247.
The Economic Times. 2009. “Amartya Sen Lauds NREGA Programme”, August
7. http://articles.economictimes.indiatimes.com/2009-08-
07/news/28486281_1_nobel-laureate-nrega-amartya-sen.
Thompson, Paul B. 2010. “Capabilities, Consequentialism, and Critical
Consciousness.” In Capabilities, Power, and Institutions : Toward a More
Critical Development Ethics, ed. Stephen Esquith and Fred Gifford, 163–
170. Pennsylvania State University Press.
Trilling, Lionel. 1948. “Manners, Morals, and the Novel.” Kenyon Review 10 (1):
11–27.
198

UNDP. 1990. Human Development Report 1990: Concept and Measurement of

Human Development. United Nations Development Programme (UNDP).
UNICEF. 2007. An Overview of Child Well-being in Rich Countries.
Ura, Dasho Karma, Sabina Alkire, and Tshoki Zangmo. 2012. GNH and the GNH
Index: A Short Guide to Gross National Happiness Index. Centre for
Bhutan Studies. http://www.ophi.org.uk/wp-content/uploads/Short-
Guide-GNH-17-May-12.pdf?cda6c1.
Valentini, Laura. 2012. “Ideal Vs. Non-ideal Theory: A Conceptual Map.”
Philosophy Compass 7 (9): 654–664.
Walsh, Vivian. 2008. “Freedom, Values and Sen: Towards a Morally Enriched
Classical Economic Theory.” Review of Political Economy 20 (2): 199–232.
Walzer, Michael. 1981. “Philosophy and Democracy.” Political Theory 9 (3): 379–
399.
Wells, Thomas. 2012. “Sen’s Capability Approach.” Internet Encyclopedia of
Philosophy. http://www.iep.utm.edu/sen-cap/.
———. 2013. “Adam Smith on Morality and Self-Interest.” In Handbook of the
Philosophical Foundations of Business Ethics, ed. Christoph Luetge.
Springer.
White, Sarah, and Séverine Deneulin. 2009. “Culture and Religion.” In An
Introduction to the Human Development and Capability Approach:
Freedom and Agency, ed. Séverine Deneulin and Lila Shahini. London:
Earthscan.
Williams, Bernard. 1989. “The Standard of Living: Interests and Capabilities.” In
The Standard of Living, ed. Geoffrey Hawthorn, by Amartya Kumar Sen,
94–112. Cambridge University Press.
———. 2007. “Realism and Moralism in Political Theory.” In In the Beginning
Was the Deed: Realism and Moralism in Political Argument, ed. Geoffrey
Hawthorn, 1–17. Princeton University Press.
Wolff, Jonathan. 1998. “Fairness, Respect, and the Egalitarian Ethos.”
Philosophy & Public Affairs 27 (2): 97–122.
Wolff, Jonathan, and Avner De-Shalit. 2007. Disadvantage. Oxford University
Press.
Yalcintas, Altug. 2009. Intellectual Paths and Pathologies: How Small Events in
Scholarly Life Accidentally Grow Big.
 199

Appendix A: Nederlandse Samenvatting*

De afgelopen 30 jaar heeft de Indiase filosoof-econoom Amartya Sen een

originele normatieve benadering voor de evaluatie van individueel en
sociaal welzijn ontwikkeld. De fundamentele vraag in deze ‘capability
approach’ is of individuen daadwerkelijk de vrijheid hebben om de
levens te leiden die zij met redenen waarderen. Deze vrijheid wordt
geëvalueerd in termen van de reële mogelijkheid (‘capability’) die
individuen hebben om combinaties van intrinsiek waardevolle doelen
(‘functionings’) te bereiken, zoals het welgevoed-zijn en het vrijuit
belijden van politieke overtuigingen. ‘Ontwikkeling’ wordt binnen deze
benadering opgevat als een uitbreiding van iemands vermogen en
daarmee als een concept dat verder gaat dan economische groei in
derdewereldlanden.
Dit proefschrift vormt een filosofische doordenking van Sens
capability benadering. In het eerste deel (hoofdstuk 1 t/m 3) presenteer
en verdedig ik mijn interpretatie van Sens werk. Ik beoordeel een
invloedrijke kritiek van politiek-filosofen die beweren dat zijn
benadering vanwege haar onvolledigheid ontoereikend is om te
discussiëren over rechtvaardigheid en daarmee geen normatieve
oordelen kan vellen over achterstelling en onrecht. Ik verdedig de
stelling dat theoretische onvolledigheid een welbewuste eigenschap is
van Sens capability benadering en geen tekortkoming. Sen is
voornamelijk uit op de evaluatie van de kwaliteit van de levens die
mensen leiden en het nagaan van hoe hier verbeteringen in aan te
brengen zijn. Zijn project beoogt dus iets anders dan een discussie over
de juiste theorie van rechtvaardigheid en stelt daarom andere eisen. Een
theorie van rechtvaardigheid is een product van de ontwikkeling en
validatie van een conceptueel denksysteem in termen van theoretische
waarden als precisie, consistentie en coherentie. De evaluatie van
welzijn is een meer empirische aangelegenheid in de traditie van de
sociale wetenschappen. De methodologische uitdaging is om een
complexe en veelzijdige casus te begrijpen door middel van het
conceptuele apparaat en de gegevens die men tot de beschikking heeft.
In het bijzonder dient de geschiktheid van de maatstaven die gebruikt
worden in de evaluatie van de casus te worden bepaald gedurende het
evaluatieproces zelf. Als theorie van rechtvaardigheid zou de capability

*
Translated by Joost Hengstmengel
200

benadering de reikwijdte van de informatie bij een dergelijke evaluatie

al bij voorbaat moeten inperken door bijvoorbeeld ‘subjectieve’ aspecten
als persoonlijke verlangens, religieuze overtuigingen and geluk buiten
beschouwing te laten. Ik analyseer Sens evaluatiekader en laat zien dat
deze de mogelijkheid biedt om beredeneerde uitspraken te doen over
achterstelling en onrecht en dat Sens verwerping van het ideaal van
theoretische volledigheid hiervoor noodzakelijk is.
In het tweede deel van het proefschrift (hoofdstuk 4 t/m 6) richt ik
mij op de mogelijke toepassingen van de capability benadering. Elk
hoofdstuk bespreekt de toepassing van de capability benadering op een
specifieke kwestie: de praktische rede (hoofdstuk 4), ontwikkeling
(hoofdstuk 5) en sociale rechtvaardigheid (hoofdstuk 6). Elk hoofdstuk
bestaat verder uit de bespreking en beantwoording van een uitdaging
aan het adres van de capability benadering. Ik begin de hoofdstukken
met de bespreking van de uitdaging (respectievelijk berusting in de
situatie, paternalisme, en het ontbreken van een politieke-filosofie over
rechtvaardigheid) en laat vervolgens zien dat de capability benadering
(met de nodige aanpassingen, maar nog steeds in lijn met Sen’s
benadering en intenties) deze uitdaging het hoofd kan bieden. Deze
exercitie identificeert en verheldert eveneens diverse toepassingen en
beperkingen van de capability benadering.
 201

Thesis summary (in English)

Over the last 30 years the Indian philosopher-economist Amartya Sen

has developed an original normative approach to the evaluation of
individual and social well-being. The foundational concern of this
‘capability approach’ is the real freedom of individuals to achieve the
kind of lives they have reason to value. This freedom is analysed in
terms of an individual’s ‘capability’ to achieve combinations of such
intrinsically valuable ‘beings and doings’ (‘functionings’) as being
sufficiently nourished and freely expressing one’s political views. In this
account, ‘development’ is conceived as the expansion of individuals’
capability, and thus as a concept that goes beyond the economic growth
of third world countries.
My thesis is a philosophical examination of Sen’s capability
approach. In the first part (chapters 1-3) I present and defend my
interpretation of Sen’s work. I examine an influential critique by political
philosophers who argue that the incompleteness of his account makes it
deficient for theorising about justice, and hence that it cannot support
making normative judgements about deprivation and injustice. I argue
that the theoretical incompleteness of Sen’s capability approach is a
deliberate feature and not a flaw. Sen is primarily concerned with the
evaluation of how well people’s lives are going, and the identification of
how improvements might be brought about. His project thus has a
different aim than debating the right theory of justice, and therefore has
different requirements. A theory of justice is a product of the
development and testing of a conceptual system in terms of theoretical
virtues such as precision, consistency, and coherence. The evaluation of
well-being is a more empirical exercise, in the tradition of social science.
Its methodological challenge is to grasp a particular complex and multi-
faceted case by means of the conceptual resources and information one
has available. In particular, the appropriate standards to be used in the
evaluation have to be determined as part of the evaluation itself.
Formulated as a theory of justice, the range of information that the
capability approach could consider in such evaluations would be limited
in advance, by for example excluding ‘subjective’ aspects like desires,
religious convictions, and happiness from consideration in any
circumstances. I analyse Sen’s methodology of evaluation and show that
it can support making reasoned judgements about deprivation and
injustice, and that Sen’s rejection of the ideal of theoretical
completeness is necessary for this.
202

In the second part of the thesis (chapters 4-6) I turn to the possible
applications of the capability approach. Each chapter considers the
application of the capability approach to a specific issue: practical
reason (4), development (5), and social justice (6). Each chapter is
organised around posing and answering a challenge faced by the
capability approach. I start the chapters by explaining a significant
challenge (respectively, acquiescence to deprivation, paternalism, and
the absence of a political philosophy account of justice), and show that
the capability approach (sometimes suitably extended but in line with
Sen’s methods and concerns) can meet each of these challenges. This
exercise identifies and clarifies several of the specific contributions and
limitations of the capability approach.
 203

Appendix B: Curriculum Vitae

Thomas Wells is British, though he has lived outside the UK for more
than half his life. He holds an MA in Philosophy and Economics from the
Erasmus Institute for Philosophy and Economics (Netherlands). He
previously studied Natural Sciences (BSc) and Philosophy (MA) at
Durham University (UK). Between his studies in Britain and the
Netherlands, he spent a year teaching English in Japan.

He is co-founder and editor of the Erasmus Journal for Philosophy and

Economics (EJPE), an international, peer-reviewed academic journal
concerned with the history, methodology, and ethics of economic
thought.

Besides his academic work, he blogs about philosophy, politics, and

economics, and he has published various essays in general interest
magazines in the area of public philosophy.